Digitized by the Internet Archive

in 2010 with funding from

University of Toronto

http://www.archive.org/details/journalc01amer

JOURNAL .b-i'Vn

OF THE

AMERICAN CERAMIC SOCIETY

A monthly journal devoted to the arts and sciences related to
the silicate industries.

Vol. 1 January, 1918 No. 1

EDITORIALS.

To the Public :

The American Ceramic Society announces the publication of
the Journal of the; American Ceramic Society, the official
organ of the Society.

The American Ceramic Society was founded in 1899, f°r the
avowed purpose of "promoting the arts and sciences connected
with Ceramics, by means of meetings, for the reading and dis-
cussion of professional papers, and by the publication of profes-
sional literature."

Up to this time, the publication of this Society has taken the
form of a volume of Transactions containing the papers presented
at the annual meeting.

At the annual meeting in February, 19 18, the Society, realizing
the need for a greater activity in the field of Ceramics, author-
ized the publication of a monthly journal to be devoted "to the
Arts and Sciences related to the Silicate Industries." This
Journal, by reason of the high standards Of the American Ceramic
Society and because of the advances which this country is making
in all that pertains to Ceramics, may well be expected to take
first place in the world's contributions to the science and tech-
nology of the industries to which it is devoted.

It is to be hoped that it may find a place in every public and
institutional library of this country, as additional service to the
reading public in general, and to technical men in particular.

Committee on Publications.

J< IURNAL OF THE

THE FUEL CURTAILMENT ORDERS.

The Ceramic industry and especially the building materials
branch lias been vitally interested in the curtailment orders
issued l>\ the U. S. Fuel Administration on April 23rd. The
orders were the direct result of the fuel shortage of last winter
when, in souk- sections of the country, plants vitally necessary to
the conduct of the war were closed down partially or completely
Tor months, on account of lack of coal.

It was quite natural that the building material industry should
be selected as the lust to be curtailed, as the entire industry con-
sumes approximately 3 2, 000, 000 tons of fuel annually, the day-
working branch consuming about one-third of this amount, or
10, 000, 000 tons. The industry being ranked as a "peace-time
essential," and not a "war essential," offered a better oppor-
tunity for a larger diversion of fuel than any other industry.
Furthermore, as it was found that for every ton of fuel consumed,
this industry puts into transportation channels approximately
three tons of raw and finished products, a total of 90,000,000
tons, a curtailment in production would result in considerable
relief to the badly overtaxed transportation system.

It was never the intention of the Conservation Division of
the U. vS. Fuel Administration to curtail all non-war industries
or to curtail any industry to a point where the demand could not
be taken care of by the industry at large. Every effort was made
to correctly estimate the demand on each industry for the year
19 1 8 and the percentages used in the orders represent the esti-
mated minimum, not the maximum. One of the principal aims
of the orders was to prevent the use of fuel in producing material
that would not be used this year, in other words, in producing
stock. It is a well-known fact that a manufacturing concern
"can do anything but shut down," but under the present condi-
tions, it was out of the question to allow a single ton of fuel to
be used unless it was absolutely necessary.

A close study of the orders with their "exception clauses"
will show how easy it is for any manufacturer to get a permit to
produce more than the minimum, provided, of course, he can
show that he has orders for the material, and provided, also,
that there is more than sufficient fuel in his district to take care

AMERICAN CERAMIC SOCIETY. 3

of the domestic demands and the war industries. These clauses
were put into the orders especially for the benefit of those indus-
tries in the "red flag" districts, which were working on govern-
ment orders, and of the plants in the western and southern sec-
tions which were not affected by a fuel shortage. Outside .of
the triangle bounded by lines drawn from Pittsburgh to Boston
and from Pittsburgh to Baltimore, known in the Fuel Adminis-
tration as the "Red Flag District," it is hardly expected that
any plant will be refused a permit to produce beyond the min-
imum specified in the orders, provided they have legitimate
orders which are acted upon favorably by the War Industries
Board.

Some criticism has been heard from members of the affected
industries because very few other industries have been cur-
tailed. There are a great number of good reasons for this. Many
industries not curtailed specifically, have been denied fuel by the
Federal Fuel Distributers, others have been under alrriost con-
tinuous embargoes issued by the Director General of Railroads,
still others, and there are many of these, have been denied raw
materials by the War Industries Board.

The cement industry may be taken as an illustration. No
curtailment order has been issued to date covering this industry,
for the reason that the cement output of the country has been
at below fifty per cent, of normal during the first five months
of this year. This was entirely due to the plants in the eastern
half of the country being denied fuel. In many cases a whole in-
dustry uses such a small amount of fuel that a curtailment order
would produce negative results, so far as fuel diversion is con-
cerned.

The' latest move of the Fuel Administration has been to
organize an "Industrial Furnace Section." As the Ceramic
industry is the largest user of fuel coming under this heading,
with the exception of the steel and coke industries, a great
amount of effort will be devoted to it by this new section. Know-
ing that there is much unnecessary waste in the use of fuel for
manufacturing purposes, the section will attempt to conserve
fuel through cutting this waste down to a minimum.

Those connected with the various industries can be of great

I |< WRNAL I '!• THE

assistance to the governmenl l>\ starting al once to look over the
plants with which they arc connected and locating any possible

leak. Uninsulated steam pipes, badly cracked kilns, poor burn

ing practice, steam leaks and poorly aligned shafting should be
especially looked after.

The section will appoint a force of field inspectors who will
inspect each plant and advise with the management. Reports
will be made to the Chief of Section with recommendations for
changes that will produce further economy. An educational

campaign will also be carried on through bulletins and the trade
magazines. Every effort will be made to bring about the desired
results without using the power conferred upon the Fuel Ad
ministration by the Lever Act.

It is naturally expected that the technical men connected with
the industries, and especially those associated with the Ameri-
can Ceramic Society will grasp the splendid opportunity offered
to assist the country in this time of stress. Many have de-
voted a great deal of time, for years past, in an effort to conserve
our fuel resources, but it has taken a great war to bring the gov-
ernment to the point where it is willing to back them up and even
to use compulsion if necessary, to bring about results. It takes
little though to foresee the benefits that will result and perma-
nentlv remain after the war is over.

THE NATIONAL RESEARCH COUNCIL.

The following Executive Order, defining the duties of the
National Research Council, was issued recently by the President:

i. In general, to stimulate research in the mathematical, physical and
biological sciences, and in the application of these sciences to engineering,
agriculture, medicine and other useful arts, with the object of increasing
knowledge, of strengthening the national defense, and of contributing in other
ways to the public welfare.

2. To survey the larger possibilities of science, to formulate comprehensive
projects of research, and to develop effective means of utilizing the scientific
and technical resources of the country for dealing with these projects.

3. To promote cooperation in research, at home and abroad, in order to
secure concentration of effort, minimize duplication, and stimulate progress;
but in all cooperative undertakings to give encouragement to individual
initiative, as fundamentally important to the advancement of science.

AMERICAN CERAMIC SOCIETY. 5

4. To serve as a means of bringing American and foreign investigators
into active cooperation with the scientific and technical services of the War
and Navy Departments and with those of the civil branches of the Govern-
ment.

5. To direct the attention of scientific and technical investigators to the
present importance of military and industrial problems in connection with the
war, and to aid in the solution of these problems by organizing specific re-
searches.

6. To gather and collate scientific and technical information at home and
abroad, in cooperation with Governmental and other agencies and to render
such information available to duly accredited persons.

Effective prosecution of the Council's work requires the cordial collabora-
tion of the scientific and technical branches of the Government, both military
and civil. To this end representatives of the Government, upon the nomina-
tion of the National Academy of Sciences, will be designated by the President
as members of the Council, as heretofore, and the heads of the departments
immediately concerned will continue to cooperate in every way that may be
required.

(Signed) Woodrow Wilson.
The White House,
May 11, 1918.

The work in ceramics, which is carried out under the auspices
of the Research Council is in the hands of three committees:
(1) the Committee on the Chemistry of Cements and Related
Building Materials, Prevost Hubbard, Chairman, office of Public
Roads and Rural Engineering, U. S. Department of Agriculture,
Washington, D. C; (2) the Committee on the Chemistry of
Glass, Arthur L. Day, Chairman, Geophysical Laboratory,
Washington, D. C; (3) the Committee on the Chemistry of
Ceramics, Edward W. Washburn, Chairman, University of Illinois,
Urbana, 111.

The last-named committee covers all branches of ceramics not
covered by the first two committees. In a future issue we expect
to present further details concerning the work of these com-
mittees.

6 J( >ik\.\i, OF THE

EDWARD ORTON, JR.

• in February 17, 1898, ten men mel in the old Monongahela
Hotel) Pittsburgh, Pa., to discuss, in an informal fashion, the
possibility of organizing a technical ceramic society. This sub-
ject had previously been discussed by Professor Edward < )rton,
Jr., and Mr. S. Geijsbeek and the time seemed ripe for such a
venture. Professor Orton was at that time Director of the only
Department of Ceramic Engineering in the United States, having
been the pioneer in the application of academic instruction to
this field of technology. He had already attained leadership in
technical ceramics and was at once looked upon as the guiding
spirit of the initial meeting and of all subsequent efforts which
led to the formal organization of the American Ceramic Society,
a year later. The task of organizing a group of men for this
purpose was by no means an easy one. The spirit of conservatism,
the clinging to cherished trade secrets, and the more or less well-
defined doubt as to the value of the technical study of ceramic
subjects, had to be overcome and men won over to the cause of
advancement. Orton's perseverance and his potent personality
attracted seventeen men willing to become charter members
of the proposed organization. His accomplishment may not
appear to have been difficult of execution, to the younger genera-
tion of today, but those who knew the conditions of twenty
year ago will appreciate the magnitude of the task. Professor
Orton had prepared for the first meeting an interesting program,
which increased the enthusiasm of the charter members and
caused others to join the Society. The meeting was a com-
plete success and the novelty of the frank discussion of technical
matters aroused wide interest.

With all the powers of his vigorous nature, the subject of this
sketch threw himself into the work of completing the organiza-
tion of establishing the rules, and of adopting a policy which has
remained as the foundation of our organization to this day. He
attracted to the Society the leading spirits of the ceramic world
and created a delightful atmosphere of good-fellowship, existing
in but few technical societies. He, nevertheless, insisted upon
continuous hard work on the part of the membership and, step

AMERICAN CERAMIC SOCIETY.

by step, he led the organization to the position it now occupies
in the industrial world.

Professor Orton has served the American Ceramic Society as
Secretary for twenty years, giving to this work his whole-hearted
devotion and making sacrifices of such magnitude that any ex-
pression of our appreciation of his services seems feeble and in-
adequate. We are so utterly in his debt.

Finally, when his country called her men to arms, Professor
Orton obeyed with characteristic promptness and is now serving
as Major in the Department of the Quartermaster General.

For a brief biography of our subject, we copy the following
paragraph from "Who's Who in America:"

"Professor Edward Orton, Jr., was born in Chester, N. Y., October 8,
1863. His parents were Dr. Edward and Mary (Jennings) Orton. Mining
Engineer, Ohio State University, 1884. On October 3, 1888, he married
Mary Princess Anderson, of Columbus, Ohio. Chemist and Superintendent
of blast furnaces, 1884-8. The first regular manufacturer of 'ferro-silicon,'
or high silicon alloy of iron, in the United States, Bessie Furnace, New Straits-
ville, Ohio, 1887-8. Entered clay industries in 1888 and managed several
plants 1 888-1 893. In 1893 he began an agitation which resulted in estab-
lishing in 1894 the first school in the United States for instruction in the
technology of the clay, glass, and cement industries, of which he was con-
tinuously, to the spring of 1916, the Director, at which time he was elected to
Research Professorship in Ceramics. Dean of College of Engineering, Ohio
State University 1902-6, and 1910-16. State Geologist of Ohio, 1899-1906.
Fellow A. A. A. S. Member of the Society of Promotion of Engineering
Education, American Society for Testing Materials, National Brick Manu-
facturers Association. Wrote: 'Clays of Ohio and Industries Established
upon Them,' Rep. Ohio Geological Survey, Vol. V, 1884,' 'The Clay Working
Industries of Ohio,' Vol, VII, 1893. Also numerous technical articles and
reports. Home: 788 East Broad St., Columbus, Ohio."

The Society, at the last meeting, voted to show its apprecia-
tion of Professor Orton's services during the past twenty years.
Realizing that mere words are inadequate, we present here the
portrait of him to whom we owe so much.

ORIGINAL PAPERS AND DISCUSSIONS.

KAOLIN IN QUEBEC.

My J. Knivi.i-:, Ottawa, Canada,

Kaolins or residual clays of any kind are rarely found in the
Arehean upland which comprises so large a portion of the Prov-
inces of Quebec and Ontario. Although feldspar dikes are fre-
quent in this area, they are mostly fresh and unaltered even
at the surface and kaolinization or even semi-kaolinization is
seldom evident. For these reasons the isolated deposit of kaolin
of workable dimensions which occurs at St. Remi is remarkable,
and so far as we know at present, is unique in the Province of
Quebec. The property is owned by the Canadian China Clay
Company and is situated near the village of St. Remi, Amherst
Township, Labelle County, about 70 miles northwest of Mon-
treal. It lies a short distance north of the magnesite deposits
described by Wilson.1

The kaolin occurs in a north and south trending drift covered
ridge, about one -half mile in width, which intervenes between
higher rocky ridges of granite and syenite gneiss and from which
it is separated by well-marked depressions. Wherever the kaolin
has been exposed in working and prospecting, the bed-rock that
contains it is the Grenville quartzite, in vertical or nearly vertical
beds, trending in a northwesterly direction. The main body of
kaolin enclosed in the quartzite beds lies along the western slope
of the ridge. The quartzite wall rock on the east of the kaolin
is mostly massive, but on the west side (Fig. 1) it is shattered
and friable for a width of 800 feet. Throughout most of the
shattered zone the quartzite is permeated by small stringers or

1 Trans. Am. Ceram. Soc, 19, 254-259 (1917).

AMERICAN CERAMIC SOCIETY.

leads of kaolin, varying from a fraction of an inch to four feet
in width. The main body of kaolin has been uncovered in
one place continuously for about 1400 feet and has been traced,
by test pits and trenches, for at least 7,000 feet. The
width varies from an extreme of about 100 feet to less than
20 feet. In a boring, kaolin was encountered to a depth of 150
ft. The covering of glacial drift varies from 2 to 10 feet in thick-
ness.

Fig. 1.
Quartzite wall rock in the kaolin mine at St. Remi, Quebec, showing ver-
tical and horizontal lines of fracture. The overburden is glacial drift.

Most of the known occurrences of kaolin are due to alteration
of the rock in situ, but the deposit at St. Remi appears to have
been deposited in its present position in the quartzite from an
extraneous source, the silica of the quartzite having been re-
placed by kaolin. The surfaces at the planes of faulting and
fracture of the quartzite rock have been channelled by the solu-
tion of the silica, and kaolin now occupies the channels. There
are large masses of mixed quartz grains and kaolin, in the main
lead, which still show the original structure of the quartzite

i,. JOURNAL I >l; THE

rock. The boundaries of the kaolin arc irregular, following the
zone of greatest fracturing in the quartzite. Where the quartzite
is dense and unbroken there is no kaolin. The quartzite con-
tained so little feldspar originally that less than one per cent,
of the kaolin could be accounted for from this source.

There are two other sources of kaolin, the one being the ortho-
clase in the beds of garnet gneiss included in the quartzite, the
other, the batholithic masses of granite and syenite which adjoin
the quartzite belt, with occasional dikes of the latter. One of
the syenite dikes in the quartzite was partially weathered, the
chemical composition being similar to that of cornish stone. The
kaolin may have been formed by rising igneous waters acting on
the feldspars of he surrounding masses and moving freely in
the fractured zone of the quartzite, the presence of tourmaline
being one of the evidences of pneumatolytic action.

There are several masses of buff and pink kaolin present in
the main lead which may owe their discoloration to the percola-
tion of water through the overlying drift. Over a large part of
the deposit, however, a snow-white clay is in direct contact with
the drift, so that the discoloration may not be all accounted for
in this way. Certain of the beds, included in the quartzite, con-
tain iron-bearing minerals, the decomposition of which probably
contributed a portion of the stain.

China Clay. — A washing plant, Fig. 2, was erected and placed
in operation in 191 2 and washed china clay has been produced
since that time. Most of the washed product is supplied to the
paper trade, although several carloads have been shipped to
potteries in the United States. The capacity of the plant has
been increased to meet the greater demand for kaolin during the
past few years. The crude material, consisting of a mixture of
quartz particles and clay, with a small percentage of finely divided
tourmaline, yields, upon washing, about 40 per cent, of china
clay. This clay is more plastic and has a greater shrinkage than
the English china clays. Its softening point is cone 34. The ex-
cellent quality of the washed clay is indicated by its chemical
composition: Silica 46.13 per cent., alumina 39.45 per cent.,
total fluxes 1 per cent., the balance being combined water.

The kaolin first mined from the deposit was uniformly white

AMERICAN CERAMIC SOCIETY.

but as the mining and stripping progressed southward, a large
amount of discolored clay was encountered. This discolored
clay contains about 2 per cent, of iron oxide and develops a light
red color upon burning and is obviously not included with the
whiter material in washing. The discolored material being
quite refractory, its utilization in the manufacture of second
quality fire brick is proposed and a small plant for this purpose
will be erected at the mine during the coming winter. Some
portions of the discolored clay contain sufficient quartzite to pro-

^ i ^rSBfJBlfiW ^ " ^^^^^^

fSSfi"-

Pig. 2.

Kaolin mine and washing plant of the Canadian China Clay Company at

St. Remi, Quebec, in 1912.

vide the necessary grog for brick making, without further addi-
tions. Additional amounts of quartzite for this purpose are
available in the crumbling west wall. The discolored clay when
washed is far more plastic than the white, in fact its working
qualities are such that it can be thrown on the potters wheel
almost as well as a stoneware clay, and may possibly be utilized
in the manufacture of novelty pottery. The discolored clay has
the open burning qualities of a kaolin, so that if a dense body,
maturing at a low temperature is desired, its mixture with a vitri-
fying clay is necessary.

[2 Jl >URNAL OF THE

Silica Products. The shattered and friable quartzite occur-
ring along the western flank <>t the deposit is permeated through

out by specks and stringers of kaolin. A washing test of a sain
pie collected over [50 feet, at right angles to the kaolin vein,
showed that 11 per cent, passed a 200-nicsh screen and that
most of this line material was kaolin. The character of this
rock suggested that it might be used in the manufacture of silica
brick of the ganister type, by simply milling the rock in a wet
pan and moulding it into brick shapes. The preliminary tests
for this purpose indicated that the behavior of the quartzite
under heating and the character of the bond produced were ex-
cellent. There seems to be no doubt that a clay-bonded silica
refractory brick, having a wide range of usefulness, may be pro-
duced from this material. The amount of raw material avail-
able for this purpose is unlimited.

It is proposed to break down the quartzite in dry pans and to
subject the ground product to a washing process which will
yield a clean quartzite sand, suitable for the manufacture of
glass and carborundum. A laboratory test for this purpose
demonstrated that most of the wall rock, when crushed, yields
a sand of the proper texture, and having the following analysis:
vSilica 9925 per cent., alumina 0.06 per cent., iron oxide 0.69
per cent. If a suitable eommercia1 process can be devised for
treating the wall rock, the utilization of that part of the deposit
for this purpose is most promising.

COMMUNICATED DISCUSSIONS.

H. Ries: The kaolin deposit of St. Remi, described by Mr.
Keele, is probably one of the most interesting in North America.
As he states, it is exceptional to find kaolin deposits within the
glaciated region, as the advance of the continental glacier scraped
off practically all residual clay that had formed in pre-glacial
times. It is conceivable, of course, that some residual material
might have remained because of the fact that it had accumulated
in rather narrow pockets, protected on both sides by hard rock.
Deposits of this type have been found at Bedford, Mass., and
near Cornwall, Conn. The latter resembles the St. Remi ma-
terial somewhat in being associated with quartzite.

AMERICAN CERAMIC SOCIETY. 13

The question of the origin of the St. Remi material, which Mr.
Keele raises, is very interesting, but at the same time it is one
that is sometimes difficult to decide. This same question has
been agitated for years regarding the Cornwall, England, china
clays. Were they formed by weathering, or by the action of
vapors or heated waters of volcanic origin rising along fractures
in the granite?

I had the opportunity of seeing these St. Remi deposits when
they were first opened up, but there was then not as much op-
portunity for studying them as Mr. Keele has since had.

It seems to me that the association of the clay with fracture
zones in the quartzite is not necessarily an indication of its deposi-
tion by rising waters, for the fracture planes might also have served
as a pathway for descending waters. On the other hand, the
apparent absence of feldspar in the fresh quartzite might be
against the theory that the kaolin had been formed by the weath-
ering of a feldspathic quartzite.

There is, however, it seems to me a third possibility, viz.,
that pegmatite solutions have invaded the quartzite in an irregu-
lar manner and that the weathering of the pegmatite deposited
has yielded the kaolin. Until we get deeper into the deposit,
definite proof may be lacking. Indeed there has always been
some doubt regarding the possibility of kaolin being formed at
depth by rising solutions of magmatic origin, a point which has
been discussed somewhat at length by W. Lindgren.1

R. R. Hice: The occurrence of this deposit is rather unique.
The whole region has been so thoroughly glaciated that we natur-
ally expect all products of decay would have disappeared, the
time elapsed since the glacial period being too short to give us
products of decay in any quantity. I am not personally familiar
with this locality, but farther west, where the underlying rock is
gneiss, cut by very frequent and often well-develope pegmatite
veins, there is no evidence of the weathering of the pegmatite
at the surface.

We do know, however, that notwithstanding the intense
glaciation which all of this portion of Canada has undergone, in-

1 "Economic Geology," 10, 89-93 (*9I5)>

, , AMERICAN CERAMIC SOCIETY.

eluding 1 1 ic- removal oi all the products of decay, and generally

of all the si'diim inn j rocks which once covered the gneiss now
exposed, there are a few occurrences still remaining of the sedi-
mentary rocks which once covered the gneiss to an unknown
distance toward the north, it is therefore clear that occasional
spots may be found where all of the products of decay have not
been eroded and this occurrence at St. Remi is probably one of
these few occurrences.

A hurried perusal of Mr. Keele's paper perhaps raises the
question as to whether or not this clay should properly be classed
as a kaolin, and again emphasizes the fact that the term "kaolin"
has been badly abused by a more or less deliberate misapplica-
tion of the term, for the purposes of industrial promotion.

SPECIAL POTS FOR THE MELTING OF OPTICAL GLASS.1

By A. V. Bleininger, Pittsburgh, Pa.

It was realized at the very initiation of the work on optical
glass, begun in the Pittsburgh Laboratory of the Bureau of Stand-
ards about three years ago, that the production of suitable pots
plays an important part in the manufacture of this type of glass.
It is evident that the pot material must resist corrosion by the
glasses, some of which are very active in this respect. The solu-
tion of the clay body in the glass causes serious difficulty through
the production of striae due to disturbance of the homogeneity,
resulting in threads or strings which are stirred into the molten
mass. It should be stated here, however, that solution of the
pot is not the only cause of striae, but that they may be due to
lack of homogeneity in the glass itself. Again, it has been ob-
served that the constituent of the pot body first dissolved by the
glass is iron oxide. The coloring power of iron, especially with
the heavy flint and barium glasses, is very intense, so that very
small quantities suffice to impart to the melt a very decided yellow
or green shade. This coloration is injurious not only for optical
reasons with reference to photographic purposes, but it is a power-
ful factor in cutting down the light transmission. Since the use
of decolorizers is out of the question in optical glass, owing to
the high absorption of light caused by them, it is obvious that the
content of iron oxide must be kept down by the use of practically
pure reagents and by the use of pots as low in this constituent
as possible. It is readily seen that, by the use of the ordinary
type of pot, the amount of iron brought into the solution might
eas ly be several times that of the original iron content of the
entire glass batch. The use of even the purest reagents would
thus be of no avail.

It would appear that pots used for this purpose should be re-
sistant to corrosion by the molten glass, yet sufficiently refrac-

1 By permission of the Director, Bureau of Standards.

t6 JOURNAL OF THE

ton to withstand the high temperature of the furnace, which

may approach 147s C, and the hydrostatic pressure- of the
liquid charge, and should be as low in iron oxide as possible.
Since the pots arc used but once, the thickness of the bottom
and walls may be cut down to the minimum. Owing to the
high value of the glass the cost of the pot, within obvious limits,
is not a serious consideration.

The first pot mixture used in this laboratory, when operations
were begun on a larger scale, was as follows:

Per cent.

Lacledc-Christy bond clay, 69-B 6

Tennessee ball clay 5

Bond clay from Anna, 111 7

Kentucky ball clay 5

"Highlands" fire clay, St. Louis 5

N. Carolina or Delaware kaolin 12

Calcined "Highlands" clay, 10-mesh 20

Calcined kaolin, through 16-mesh 40

In using this mixture it is necessary to calcine the kaolin to a
temperature of not less than the softening point of cone 14. This,
of course, increased the cost of the pots. Later, the calcined
kaolin was replaced by a mixture of 80 per cent, kaolin (N. Caro-
lina, Florida and Georgia in equal parts), 10 per cent, flint, and
10 per cent, of feldspar, fired also to cone 14. This gave a sharper
and less friable grog. Both compositions gave excellent results
with reference to the standing-up qualities and the resistance
of the pots to corrosion. Still, the iron content of the mixture
was somewhat too high and it was thought desirable to work
out a composition akin to that of a hard porcelain. After some
experiments the following composition was arrived at:

Per cent.

Calcine 43

Kaolin 25

Plastic bond clays 25

Feldspar 7

AMERICAN CERAMIC SOCIETY. 17

Here, the calcine consisted of the kaolin, feldspar and flint mix-
ture mentioned above. It is evident that in blending such a
number of materials great care is necessary in order to secure
thorough dry mixing and pugging. These pots likewise gave
very satisfactory results.

The question of procuring the calcine in larger quantities
was found to be an annoying one and we cast about for a cheaper
source of grog. This was found in the waste bisque of wrhite
ware potteries, which is obtainable at a reasonable price and in
sufficiently large quantities. This type of body corresponds to
the general composition of 35 per cent, kaolin, 15 per cent, ball
clay, 14 per cent, feldspar, and 36 per cent, flint. The white
granite bisque softens at about cone 30 and possesses, on account
of the high flint content, excellent standing-up qualities under
conditions of pressure at high temperatures. It is evident that
porcelain bisque, whether table ware, electrical porcelain, or floor
tile, would not answer for this purpose, owing to the low refrac-
toriness. There would be no objection, of course, to the use of
this kind of material in replacing the feldspar introduced, the
proportions of white ware bisque and porcelain bisque introduced
being such as to maintain the desired refractoriness. In employ-
ing white ware bisque, it should be obtained in as clean a condi-
tion as possible and care should be taken not to produce too large
a percentage of fines (finer than 80 mesh) in grinding. If this
grog should become too fine, it might cause serious cracking of
the pots during cooling, probably at a temperature of about
625 ° C when the glass is still in a semi-fluid condition. This defect
is due to vitrification, facilitated by the presence of the fine
particles of grog and the resulting homogeneous porcelain
structure. In satisfactory pots, the grains of grog should still
retain their identity as shown by their outlines.

At the present time both white ware bisque and old pot shell
(from porcelain type pots) are being used and the composition
of the pot batch is as follows:

is JOURNAL ' »r THE

White ware bisque, through to mesh 35

Po1 shell, through i" mesh n,

Feldspar 3

Flint 4

Tennessee bali clay No. 5 15

Illinois bond clay 5

kaolin 28

100

The ground and screened grog and other components of the
body batch are weighed, mixed, and then tempered in a wet
pan. It is desirable, in tempering, that the mullers be raised
about y4 inch off the bottom of the pan so that the grinding
action is reduced to the minimum. The clay is then passed
through a vertical pug mill and stored as long as possible. The
body should be aged at least a month but under our conditions
this has not been possible, owing to the demand for pots. The
treading of the plastic body has been done away with as unneces-
sary. The pots are built up by hand in the usual manner and
both the open and covered types are made. The pot used in
our laboratory is 34 inches in diameter (outside), 27 inches high,
and has a bottom 4 inches thick. The wall tapers from 2l/i
inches (top) to 3V2 inches (bottom). The weight of each pot
is about 500 pounds. After drying for about four weeks, the
pots are ready for the furnace.

One or two pots each week are also being made by the casting
process. The composition of this body is as follows:

Per cent.

White ware bisque 48

Plastic bond clay 23

Kaolin 24

Feldspar 5

The slip used carries 80 per cent, of solids and hence 20 per
cent, of water. A mixture of equal parts of sodium silicate and
sodium carbonate is used as the electrolyte and the amount
added, in terms of the dry weight of the body, is 0.20 per cent.

AMERICAN CKRAMIC SOCIETY.

19

The arrangement of the mold, as shown in the illustration of Fig.
1, needs no detailed description. It is, of course, necessary to
hold the core firmly in position by means of cross beams and tie
rods as the upward hydrostatic pressure of the slip is very con-

Fig I.
Casting mold with core removed.

siderable. The casting, as now practiced, is through three fun-
nels arranged symmetrically around the circumference of the
mold. The liquid in the funnels is replenished as the level is
lowered. During the last stage of the casting, the settling is

|< HJRNAL I 'l; THE

very slow, so thai the- funnels may be filled and the process allowed

to go Oil without close supervision. Absorption of the liquid
eeases alter about 10 hours and the core ma\ then be removed.
The outer mold may be removed alter 24 hours. Suitable hoists
are necessary for the handling of the mold parts and the pots.

The process of preparation of the slip is very simple, consisting
in introducing the weighed, ground body constituents into a
double blunger together with the required amount of water in

Fig. 2.
The first set of porcelain pots made at the Pittsburgh Laboratories, Bureau

of Standards.

which the electrolytes have been dissolved. When the desired
consistency has been attained, the slip is allowed to run into a
large bucket which is conveyed to the mold, hoisted and the
slip discharged into the funnels. No difficulty has been ex-
perienced in drying the east pots in less than 3 weeks. They
have given very good satisfaction in the glass furnaces and, if
anything, are superior to the hand-made ones, especially on ac-
count of the fact that they have thinner walls, the thickness

AMERICAN CERAMIC SOCIETY. 21

varying from 2 to 3 inches. For this reason less time is required
in heating up.

In the use of porcelain type pots it is important to raise the
temperature of the furnace to not less than 14000 C before the
charge is introduced, in order that the body may 'become vitri-
fied. The temperature is then lowered and the glass batch is
introduced into the pot. Unless this is done, the benefits of the
porcelain structure are lost and the pot may corrode as much,
or more, 1han one of the ordinary type.

The same body composition is also used in the casting of the
stirring rod parts, which are immersed in the glass during the
mixing stage of the melt.

( )ur experiences with pots for the melting of optical glass have
been freely imparted to pot manufacturers and makers of fine
glass and, as a result, pots of similar composition are being pro-
duced by several firms.

The illustration (Fig. 2) shows the first large porcelain pots
made in this laboratory, in the fall of 191 7. Small porcelain
pots, holding about 30 pounds of glass, were first used by us in
May of the same year.

COMMUNICATED DISCUSSIONS.

C. H. Kerr: I have been exceedingly interested in reading
this paper, my chief thought being one of regret that so few
readers have been through the mill so as to have the correct
perspective to appreciate the work presented. Nothing, but
the actual experience of working on and solving (more or less
completely) the problem of making a pot that permits the pro-
duction of real optical glass, can give one a proper realization
of the great difficulties involved. Very few people, even among
those who have had some contact with optical matters, realize
how absolutely we depend upon optical glass in this great war.

Practically all of the materials used in all industries offer some
possibility of substitution. It is frequently difficult, in the use
of structural materials, to obtain results by the substitution of
materials that are fully equal to the original, but almost always
an approximate substitution can be made. This is not true

JOURNAL OF THE

in the optical glass work. We niusi have the optical glass for
fire control instruments for our Army and Navy and nothing will
take its place.

It is very fortunate, although not generally realized by tin
American people, that we now have an optical glass industry
in this country well started and we are able to take care of our
military requirements with a very considerable degree of satis-
faction. Contrary to usual belief, the production of optical
glass is not, at the present time, the limiting feature in the pro-
duction of optical instruments for fire control purposes.

Throughout the development of the optical glass industry
the manufacture of a pot in which it is possible to produce a good
grade of optical glass has always been of the first importance
and, very frequently, it has been the limiting feature in a par-
ticular line of work. For the production of most of the types
of optical glass it has been very difficult to secure clays of the
requisite purity and composition and which, at the same time,
possess the required working properties necessary to the produc-
tion of full-sized pots satisfactory for use in the usual optical
glass procedure. It is not difficult to get the required purity,
or the required composition, but to combine therewith the physi-
cal properties of the clay or mixture which are necessary, has
been exceedingly difficult. Such clays and mixtures, however,
have been found and are being used successfully. So far as the
question of general composition is concerned, my experience
has been somewhat parallel with that cited in the paper and
has demonstrated that compositions of the types indicated pro-
duce excellent pot batches.

In optical glass pot making it is especially true, as stated in
the paper, that an excess of fine material in the grog is a very
frequent source of disaster. This trouble is best overcome
by a control of the non-plastic part of the mixture to see that
it has previously been burned to a sufficiently high temperature,
so that, after grinding, the resulting mass is a comparatively
coarse material. If the material is too soft it very easily breaks
down into a fine powder.

The casting process undoubtedly has a great future, especially
in the manufacture of pots of this type. Probably the most

AMERICAN CERAMIC SOCIETY. 23

important advantage lies in the fact that a denser structure
can be obtained by the casting process. The most important
application of this point is in the fact that a mixture of a density
such that it is hardly practical for use by the ordinary hand-form-
ing process, can be worked very satisfactorily by the casting
process. Along this same line, it is possible to increase the
rapidity and safety of the drying process when using cast pots,
which has the double advantage of being commercially superior
for an ordinary batch and also making commercial a batch which,
by the usual process, is hardly capable of commercial handling.
Another very great point in favor of the casting is the great in-
crease in homogeneity of the mixture to a degree which is not at
all possible by the usual hand-forming method.

E. W. Tillotson: Dr. Bleininger has brought out one point
that may well be given consideration by all users of glass pots.
He states that a pot of the porcelain type, to be efficient, must
be burned at a temperature considerably above that of the glass
furnace. Such a procedure is rarely or never followed in a glass
factory. In fact, the pot arch seldom reaches furnace tem-
perature and the result is incomplete burning, especially of the
bottom of the pot.

It would appear that the glass maker should give serious at-
tention to the revision of the construction of the pot arch for the
purpose of securing a more even burning of the pot and in order
that, by burning at a higher temperature, a pot may acquire its
maximum density. Such a procedure may be expected not only
to lengthen the life of a pot but also to improve its service as re-
gards corrosion, etc.

C. E. Fulton: Mr. Bleininger has covered the subject very
thoroughly and there does not seem to be anything of value which
[ can add in the form of discussion. We have gone through prac-
tically the same experience and while the composition of our
optical pot batch is somewhat different from that developed by
the Bureau, it is essentially the same type, the chief problem, of
course, being to secure a pot with a very dense structure and con-
taining a minimum amount of iron. We contemplate making

•i AMERICAN CERAMIC SOCIETY.

some pots following the formula given by Mr. Bleininger, in
order to see how they compare with the pots we arc now using.

I). W. R»>ss: A matter of prime importance iii connection
with this paper is that glass pots made in accordance with the
lines therein laid down are today being used in practically all
, the optical glass plants of the country and are producing the de-
sired results. Thus, a great burden of experimenting has been
cleared away from the optical glass manufacturer and the pot
manufacturer alike.

The compositions, as given for both fire clay and semi-porce-
lain pots, should form a foundation from which other branches
of the glass industry may work out pots, etc., suitable for their
special needs.

C. R. Peregrine: The Society is fortunate in having papers
like this by Mr. Bleininger contributed to its Journal.

It would be interesting to know howT much corrosion takes
place when the pots have been properly preheated, and whether
this is greater for the cast pots than for those built up by hand
in the usual way.

The policy of the Bureau of Standards, through its Director,
which makes its experience available to manufacturers, points
toward a new era in which cooperation will do much to place
American industries in the front rank.

A. V. Bleininger: I believe there is nothing to be added
to the above discussion. The question of Mr. Peregrine is al-
ready answered in the text.

THE EFFECT OF GRAVITATION UPON THE DRYING
OF CERAMIC WARE.

By Edward W. Washburn, Urbana, Illinois.

That the force of gravitation is able, under certain conditions,
to appreciably influence the drying of a mass of clay, was called
to the writer's attention as a result of the following experience :

A ball of moist, plastic clay about four inches in diameter was
placed on the shelf of an ordinary chemist's desiccator, the lower
part of which was filled with water, the idea being to keep the
air surrounding the ball of clay moist so that the clay would
not dry out on standing overnight. It happened that the clay
was left in the desiccator for a month, and at the end of this time
it was found to have dried out so completely that it readily broke
into pieces when dropped upon the floor. On examination
even the central portions of the ball were found to be hard and
dry.

At first sight, this behavior seemed rather astonishing, as the
ball of clay was standing in an air-tight vessel only three inches
above the surface of the water and the air surrounding the clay
should, it would seem, have been almost saturated with mois-
ture. In seeking the explanation for this behavior the writer
came to the conclusion that the water had been actually drawn
out of the clay mass by the attraction of the earth. In other
words, the clay must have dried for the same reason that water
runs down hill.

In order to test the conclusion that the force of gravitation
was responsible for the drying, the experiment was repeated
under conditions of more exact control. For this purpose a 10-
gram mass of moist clay (containing 30 per cent, of water) was
suspended by means of a wire from the bottom of the rubber
stopper of a bottle containing a small quantity of water, the
height of the clay above the surface of the water being about 6
centimeters. (See Fig. ia.) The bottle and its contents were

>6

fOURNAX OF THE

then completely immersed in a large body of water so as to in-
sure constant and uniform temperature conditions on all sidea
of the bottle. Prom time to time the mass of day was removed
from the bottle and weighed. The experiment was continued
for eleven days, during which time the muss of clay continuously
lost weight at the rate of about 30 milligrams per day. At the
end of the eleventh day the mass of clay had lost a total of 0

a

t>

Fig.

gram of water. This water must obviously have distilled out
of the clay, under the influence of the attraction of the earth,
and condensed into the water in the lower part of the bottle.

To make certain that this was the correct explanation, the
mass of clay at the end of the eleventh day was lowered so as
to be on the same level as the water in the bottom part of the
bottle, but without touching this water, the arrangement being

AMERICAN CERAMIC SOCIETY.

27

that shown in Fig. 16. The experiment was then continued and
the mass of clay was found to continuously increase in weight,
showing that water was now distilling from the liquid water in
the bottle, and condensing into the mass of clay. The distilla-
tion in this case was, of course, due to the fact that, owing to the
presence of small quantities of soluble materials in the clay, the
tTapor pressure of the water in the clay was lower than that of
Dure water. The experiment was carried out in duplicate and
:he two curves showing the variation of the mass of the clay

C20

7

/

/

0.10

got

A

/r.

V5

1

a

G

t>

%0.I0

I

A

b

/

!v

020

V

\

/

*

\

1

030

Time of fnposure to Masf Air (Days!

Fig. 2.

from day to day are given in Fig. 2. A repetition of the experi-
ment using a small glass capsule filled with water, in place of
the clay cylinder, yielded a similar result.

The only unexpected feature of this experiment was the com-
paratively great rate at which the water was removed from the
:lay, under these conditions. Other things being equal, we
?hould expect this rate to be determined by the difference between
the vapor pressure of the water in the clay and the partial vapor

28 JOURNAL OF THE

pressure of water in the surrounding air. This latin pressure
is greatest, of course, at the surface of the liquid water in the bo1
torn of the bottle and decreases steadily with the distance above
this surface. The relationship between this latter pressure and
the height above the surface of the water is expressed by the
equation

i\p = gDdh (i)

or lor short distance's, with sufficient accuracy, by the equation

A/> = gDh, (2)

where A/-> is the difference in the partial pressure of the water
at the surface of the liquid water and at the height h above this
surface, D is the density of saturated w^ater vapor at the tem-
perature in question and g is the acceleration due to gravity.
For a temperature of 20 ° C and a height of 6 centimeters we find
upon calculation that

Ap = 980 X 17 X io~6 X 6 = o. 1 bar. = about 0.000003 inch
of mercury, which is obviously a very small pressure difference.

Aside from the purely theoretical interest which attaches to
the above experiment and conclusion, it is perhaps worth while
to point out that in order to keep a mass of moist clay from dry-
ing, by inclosing it in a chamber containing water, the water
should not be placed at a lower level than the mass of clay which
is to be kept moist. Moreover, there would appear to be no rea-
son why the force of gravitation should not be made use of in
some cases to control the extent and rate of drying of ceramic ware
since, at a given temperature, both of these factors are deter-
mined by the perpendicular distance between the piece of ware
and the bottom of the containing chamber, provided that the
air in the chamber is not stirred. Stirring the air would ob-
viously retard the rate of gravity drying but would of course not
stop it entirely unless the stirring were 100 per cent, efficient.

It is also interesting to note that since according to Equation
2 the value of p is proportional to D, the density of the saturated
vapor, it follows that the rate of gravity drying would increase
with the temperature. That is, a mass of moist clay placed in
a gravity dryer filled with saturated steam at ioo°C would dry

AMERICAN CERAMIC SOCIETY. 29

more rapidly than in the same dryer at o° C, although at the latter
temperature the air in the dryer would contain very little mois-
ture. In other words, the rate of drying in a gravity dryer is
letermined by the vapor-pressure gradient in the dryer and
lot by the dryness of the air in the dryer. Indeed, as we have
>een, the more moisture the air of such a dryer contains, the
nore rapid will the drying be, a conclusion which at first sight
ippears paradoxical.

In this connection attention may also be called to the fact that,
n operating a moist closet for storing prepared clay, the drying
iction of gravity upon this clay cannot be eliminated by hanging
Aet cloths along the sides of the closet, provided the action of
capillarity is depended upon to keep these cloths wet; that is,
Provided they are supposed to be kept moist by having the lower
aids dipping into water in the bottom of the closet. It is true
:hat in such a manner the water placed in the lower part of such
i closet may be lifted to the same level as the clay, but such water
i. c, water held up by capillary force acting against gravity)
will be under tension and the effect of tension upon a liquid is
ilwavs to lower its vapor pressure. In fact the vapor pressure
it the surface of the meniscus in a capillary- tube (in which equi-
librium has been established between gravity and surface ten-
sion; will be identical with the vapor pressure at the surface of
he water outside of the capillary- tube, although this latter sur-
race will be at a lower level than the surface inside the capillary,
rhis statement will be true regardless of the size of the capil-
arv; that is, regardless of how great the difference in the two
levels inside and outside may be.

Although not bearing directly upon the subject of this paper,
t is perhaps worth while pointing out in this connection that
a method for the direct measurement of the osmotic pressure
)f any very dilute solution may be based upon Equation 1 . The
method requires no membrane of any kind and consists simply
n finding, by experiment, at what height above the surface of
the pure solvent a small quantity of the solution must be placed
in order that it shall neither gain nor lose weight by gravity
listillation. This height is identical with the height of the osmotic

30 JOURNAL OF THE

column, as can be readily demonstrated l>v purely thermodynamic
reasoning. The method would, of course, in practice, be restricted

to solutions of such dilution that the height of the osmotic col-
umn would not be inconveniently great. It seems quite within
the bounds of possibility thai an accurate technique could be

worked out for applying this method, which would have many
obvious advantages over the other colligative property methods

in the case of very dilute solutions.

In addition to the cases discussed above it seems probable
that the influence of gravity upon vapor pressure also plays an
important part in some of the phenomena connected with the
swelling of colloidal gels. Thus it has been observed1 that if a
gel, which has been soaked in water until the swelling has reached
its maximum, be then raised out of the water, the gel loses most
of its water and consequently shrinks greatly in volume. The
water lost in this way must evidently have distilled out of the
gel under the influence of some force acting upon the water in
the pores of the gel but not upon the liquid water below it. In
view of the experiments recorded in this paper, it seems highly
probable that the force of gravitation is responsible for this be-
havior. It is also probable that some experimental results ob-
tained by Foote2 are likewise the result of the influence of gravi-
tation.

DISCUSSION.

Prof. Watts: I would like to ask Dr. Washburn if the two
clays were the same or different clays.

Dr. Washburn: The same clay.

Prof. Watts: Have experiments been conducted to show
the difference in the rates at which very dense or open clays

1 Schroeder, Proc, Amsterdam Acad., 15, 1078 (1913); 17, 92 (1914).
The so-called Schroeder's paradox was further studied by Wolff and Buchner
(Z. physik. Chetn., 89, 271 (1912)) who found, when precautions were taken
to conduct the experiment so that gravitation could not act differently upon
the water and upon the swollen jel, that under these conditions no difference
between the vapor pressures of the two could be detected.

2 Foote, J. Am. Chem. Soc, 30, 1390 (1908).

AMERICAN CERAMIC SOCIETY. 31

would dry? Do you consider the rate of drying a matter of
the size of grain of the clay?

Dr. Washburn: You can make the rate anything you like
by regulating the height at which you place the clay above the
water.

Prof. Watts: This subject brings to my mind a very inter-
esting point in connection with the potter's wet closet. The
cup maker has a room in which he keeps his cups subject to jig-
gering and prior to turning them, the cups remaining in the closet
almost indefinitely. They are placed in the closet as soon as
removed from the mold and remain until ready for turning.
The closet is an ordinary closed chamber, having a large pail of
water at the bottom. It is significant that the cups may remain
in the closet for a week or ten days, or even longer, before being
turned, and yet they appear to be in satisfactory shape for turn-
ing and I do not know of any particular reason why the mois-
ture should remain in those cups, unless there is some circulation
of the air in the closet. The cup closets are usually chambers
built in the side of a large, open room which is normally warmer
near the ceiling than near the floor. We may presume that the
increased temperature at the top of the room, outside the cham-
ber, develops a warmer zone inside the chamber and maintains
circulation of moist air inside the chamber, keeping the cups
moist. A potter, in reading this paper of Dr. Washburn's,
would immediatelv raise this question and I believe that my
explanation is correct.

Mr. Linbarger: We have conducted a number of experi-
ments on humidity drying systems and find that, when using
as high a relative humidity as 90 per cent, when the air in the
dryer was not in motion, we get a loss in weight of the material
in the dryer. This loss takes place at a definite rate, for when
we place the material on an accurately calibrated balance and
take frequent readings from the outside, so as not to disturb
the air in the dryer, we are able to measure the rate of evapora-
tion. We note, in this connection, that with all other conditions
constant, we have more evaporation when we are using a 90

JOURNAL OF THE

pei cent, relative humidity than when the relative humidity is
65 per cent. We have been at a loss to account for this ph<
nomena, bul I think Dr. Washburn's paper presents a very sati
factory explanation.

Mr. Riddle: I would like- to ask I'rof. Watts if the same con-
ditions are noted in case the walls and shelves of the drying
closets are covered with plaster which is kept in a moistened con-
dition? Would the use of plaster materially alter the moisture
conditions in the closet?

Prof. Watts: I was referring to the ordinary wooden con-
struction for the closets. I was not referring to closets having
plaster coated walls or shelves.

Mr. Riddle: I would like to ask Dr. Washburn: under high
humidity conditions, is any of the drying due to capillar}- action?
For example, in a tunnel dryer for brick, if the humidity is de-
creased too rapidly, the flow of moisture from the inside to the
outside of the ware is broken.

Dr. Washburn: In the dryer I speak of, the drying is all
due to gravity. The water in the clay gets to the outside by
capillar}' forces to a great extent, but the drying itself is due to
gravitv. Air. Watts did not state that the cups did not dry,
that is they were not weighed. This action, of course, is a com-
parative!)' small one and I think if those cups were weighed, they
would have been found to have actually lost water; the}' might
not have lost enough but that they could be turned satisfactorily-
but they actually would lose water if the air were not stirred.

Mr. Barringer: I think this paper touches upon funda-
mental questions in connection with drying, and in discussing
them we should be very careful not to lose sight of the proper re-
lation of the various, factors affecting drying. One of the gentle-
men spoke of "humidity bringing about quick drying or the loss
of moisture." Undoubtedly, the high humidity was applied
with a certain temperature, and it i£ practically impossible in
drving to separate the various elements concerned.

AMERICAN CERAMIC SOCIETY. 33

As I understand it, the theory of drying involves controlling
he relation of the vapor pressure of the water in the ware and
he pressure of the vapor of the air, outside the ware. There are
actors affecting both of these. For instance, the vapor pressure
f the water in the ware increases as the temperature, and the
ressure of the vapor outside (the moist air) increases as the
.umidity. The first step in drying is to balance these pressures,
'here is another factor that must be considered in drying, and
hat is, the heat conductivity of the clay wares themselves.
rery often when we speak of high temperature, we do not mean
he temperature of the clay ware itself, but the temperature of
tie air.

It seems to me that this element of the mechanical motion of
lie water ought to be considered in connection with the theory
f drying and the fundamental considerations I have mentioned,
would like to ask Dr. Washburn whether, under this drying by
ravity, he would classify centrifugal force? Suppose you should
Dtate a piece very rapidly; would the water not tend to leave
lie clay upon the same principle as by gravity ?

Dr. Washburn: The rotation would tend to increase the
ressure of the water on the periphery; an increase of pressure
n a liquid always increases the vapor pressure of that liquid.

Mr. Davenport: I would like to ask Dr. Washburn if he
ad an increasing concentration of moisture in the lower portion
f the sample?

Dr. Washburn: No, the sample was very small, I did not
ttempt to detect it, but of course the upper part must have
ried faster than the lower part.

Prof. Watts: I would like to ask Dr. Washburn if the me-
lanical operation of removing this piece from. the atmosphere
1 the bottle, and weighing, could have had any material effect
3 regards the evaporation of the moisture that might have been
urried to the surface of the piece?

Dr. Washburn: In answer to Mr. Watts' question, I will
iv 110, not as far as the correctness of the final result is concerned;

34 AMERICAN ceramic society.

the effects he mentions would merely produce slight irregulari
ties in the points along the curves, bu1 when the clay was lowered

to the surface of the water and began to increase iu weight, it
showed that the effect you mentioned could not have vitiated
tin conclusions in ahv wav.

TEST OF A PRODUCER GAS-FIRED PERIODIC KILN.1

By C. B. Harrop, Columbus, Ohio.

I. Introduction.

The use of producer gas in connection with continuous ceramic
kilns has undoubtedly passed through the purely experimental
stage. Suggestions that producer gas be used in individual
periodic kilns was at first looked upon with considerable suspicion
and doubt and some of our recognized engineers even stated
freely that the plan was not practical, offering as their reason the
belief that producer gas, being a comparatively lean gas and low
in calorific value, could not be used for high temperature work,
except that hot air be used to support combustion.

One of the first practical attempts to use producer gas, in
connection with periodic kilns, was made at the plant of the
U. S. Brick Company, at Tell City, Indiana, where a system was
installed in 1914. This plant is operating continuously and from
a commercial standpoint, the installation has proven entirely
satisfactory to the owners. Other plants have since been put
into operation, not always successful in every particular, but at
least proving that with suitable coal, properly selected and in-
stalled equipment and intelligent management, the plan is prac-
tical for wares requiring temperatures of 11500 C or lower.
To how much higher temperatures it would be practicable to go,
tests made so far throw no definite light.

II. The Installation.
(a) Producer. — The producer is of the round, stationary
pressure type, with a double system of blast piping, delivering
steam and air to the bosh as well as to the center tuyere (Fig. 1).
The inside diameter of the producer is io' 6" and it has a maximum
gasification rate of 1250 lbs. of coal per hour. The steam is
supplied by the main boiler plant of the factory and is carried
it 25 lbs. (gage pressure) at the blower.

1 By permission of the Director, Bureau of Standards.

36

l< >URNAL OF THE

(6) Fuel.— The fuel used was an Indiana coal known a^
"Washed No. 3 Ayrshire Nut" and was charged at intervals
averaging 25 minutes throughout the test.

Pig.

-Gas Producer.

Each charge of coal was sampled and two combined samples
were retained for examination, one covering each half of the
period of the test. The two analyses were averaged on the basis
of the relative amounts of coal charged during the two halves,

AMERICAN CERAMIC SOCIETY.

37

hese amounts being indicated in Fig. 2, together with the heating
rallies. The average idtimate analysis was as follows:

H ydrogen 6 . 10

Carbon 66 . 3 1

Nitrogen 1.52

Oxygen 20 . 59

Sulphur o . 90

Ash 4.58

(c) Gas Flues. — The main gas flue leading from the producer
oward the kilns is of brick, with arched top. Its original size
/as 5' wide by 5' high, but it was found that at the place where

ztoo

/iv/c

*? rent.

7 S%/j/>

??7Sfy

*

0/7SCS/TT
Si

'ofay*

'7*7 rf'J

<*l GO,

1
^J — i

/
1

~^>— "

1 \
i \

/

1

\

\

\ J

\

\

A^

,

V

\

\
\ 1

c''

,S

v

!>

\

~*/'

/

■*j

V

X wo

fl'

1
1

\

1
1

1

N

JOO

\ .'

V

1

O <* /* /0 *4 JO Jt f* -98 J? 60 <*<» 7i Tg *V SO 36

Fig. 2. — Fuel charged into producer over 6-hour periods.

he gas samples were taken, the flue was only 4' deep, evidently
ue to a deposit of tar and soot in the bottom. There is a thin
iver of sand and about 12" of earth fill on top of this main flue.

Two branch flues, 24" wide by 26" high, partly encircle the kiln
nder test. The burners are connected with these branch flues
y laterals consisting of 10" sewer pipe, as indicated in Fig. 3.

(d) Air Supply. — Air for combustion is supplied by a No. 9
Dp angular discharge centrifugal blower, manufactured by the
lational Blower Company, of Milwaukee, Wis. This blower is

38

Jl HJRNAL OF THE

installed in an isolated building and is belt-driven from a line
shaft, which in turn, is driven from a 7" X 12" slide- valve throt-
ling engine. The speed of the blower was 845 r. p. m. at the

beginning of the test and was finally increased to 1065 r. p. m.

(e) Air Ducts. — The air blower discharges into an overhead
metal pipe of No. 22 gauge steel, 16" in diameter. From this pipe
a branch or bustle pipe, 12" in diameter, practically encircles
the kiln about on a level with the top of the kiln wall. Con-

AMERICAN CERAMIC SOCIETY.

39

necting each burner with this bustle pipe is a 4" diameter down-
take fitted with a slide damper.

(/) Kiln. — The kiln used for this test, which is of the round
down draft type (Fig. 4) and designated as No. 1, is the nearest
3ne to the producer. This was not according to design, but it
happened to be the kiln set at the time that the Bureau rep-
-esentatives were free to carry on the test.

This kiln is 30' in diameter, is y' from the floor to the spring
3f crown and the crown has a 7' rise. The crown is 12" thick and

1. , OF '•i^&mggk&

Fig. 4.— The kiln.

:he side wall 22" thick. The hob is 36" thick and stands about
50" high above the ground. There are eight furnaces provided
with rectangular bags, 18" X 30" inside and 5' 3" high. The
din has two doors, each 4' wide by 5' 6", high.

The floor of the kiln is perforated over its entire area and is
served by the rather common cross-head-flue draft collecting
system.

40

JOURNAL OF THE

The tnaill draft Hue, the top Of which is about 2' below tin

ground line, is 36" square inside and leads to a 4-compart

ment chimney, 40' high, situated about 10' from the kiln. Each
compartment of the chimney is 36" square and serves but one
kiln.

The cast iron burners (Fig. 5) which are of a patented «i
are supplied from the underground (lues with gas, which passes
directly into the central chamber of the burner. For regulating

Fig. 5. — -Gas Burner.

the gas, there is supplied a flat sliding damper, resting on the
bottom of the central chamber, and with a handle for adjusting,
which passes through the hinged outer end of the burner. The air
for combustion, coming through the downtake, enters an annular
space ahead of the gas inlet and surrounding the central chamber,
the gas and air not mixing until they leave the inner end of the
burner.

The underside of the furnace arch, in the main wall, rises 4"

AMERICAN CERAMIC SOCIETY. 41

ibove the top of the kiln hob, thus leaving an opening 4" high In
ibout 15" long into the furnace chamber.

The kiln is started with four fires only, which continue for the
Irst 12 hours and one full size brick is laid across each of the
openings above the hob. At the end of the first 12 hours, the
'emaining four burners are lighted. In 18 hours after starting
:wo bricks are laid across the openings above the hob and in 30
lours these openings are daubed up tight.

(g) Ware. — The product of this plant is common building
3rick and face brick of the rough or texture variety, the stiff -mud
Drocess being employed for both. The raw material is a rather
;andy alluvial clay from the immediate vicinity, bordering the
)hio River. This clay, under oxidizing conditions, burns to a
ieep red color at about cone 02 in the bottom of the kiln. Under
educing conditions, however, the color passes through various
shades of tan, brown and finally to gun metal.

Three courses of common (smooth) bricks are set on edge on
:he bottom and one course on the fiat. On top of these are 44
courses of face brick set fiat, making a height in the center of the
ciln of about io', unburned. The setting is 3 on 2, in 2-brick
benches.

Three and a half to four days is the usual time required to burn
me of these kilns. The method followed is to increase the tem-
jerature to about 995 ° C (as indicated by an electric pyrometer,
:he couple of which hangs through the center of the crown) and
;hen to soak for 24 hours at a slowly increasing temperature.
Arhile pyrometrie cones are placed in the setting near the wicket,
he time of finishing is determined by measuring the settle through
he center of the crown. The first settle measurement is taken at
ibout 72 hours and the gas is shut off as soon as the settle totals
[ 1 inches.

III. The Investigation.

(a) Object. — Considering the doubt as to the feasibility of
;his method of burning clay wares that has existed in the minds
)f many who have been interested in the reduction of kiln burning
expense and labor, it was believed that an investigation of the
ystem would show whether or not it embraced any real points of
nerit.

42 JOURNAL OF THE

When compared with results <>1" similar tests made on solid
fuel-fired, natural gas-fired or oil-fired kilns, and taking labor
into account, it can be seen wherein and to what amount, one has
an advantage over the other.

It was not desired to make a test of the particular producer
in use, but as the producer is a part of the system, it was necessary
that the investigation include both producer and kiln. In
analyzing the results, we can consider the system as a whole, or,
we can consider merely the kiln as being supplied with gaseous
fuel of a certain composition and at a certain temperature, and
study its performance under the existing conditions.

(b) Apparatus. — A Hygrodeik was used for determining the
temperature and humidity of the atmosphere.

A platform scale was used for weighing the coal charged.

A train, consisting of a water-jacketed glass tube and an asbes-
tos packed glass filter,1 was used for sampling the tar and soot
from the main gas flue. The gas carrying these ingredients was
drawn through the train into a water aspirator bottle. The
drawing out of the water from this bottle, between two marked
levels, indicated the amount of gas drawn through the train. A
thermometer was inserted in the aspirator bottle, the readings of
which permitted a correction of volume to a basis of the tem-
perature in the gas main.

For determining the water vapor in the producer gas in the
main flue, a glass tube '/2" inside diameter by 18" long was used.
This tube was filled three-fourths full of calcium chloride and in
the remaining space was placed phosphorus pentoxide. Known
quantities of gas were drawn through this tube (passing through
the calcium chloride first) in the same manner as in sampling the
tar and soot.

An iron pipe sampling tube, having small holes at 4" centers,
was used for sampling the producer gas over the entire depth
of the main. Gas was drawn continuously from this sampler
through a bubble bottle by a wrater aspirator. A 6-lb. glass
sample bottle was used for drawing off samples of gas for analysis,
just before passing through the bubble bottle.

A Williams apparatus was used for the analysis of the pro-
1 Bur. Mines Tech. Paper, 137, p. 20.

AMERICAN CERAMIC SOCIETY. 43

ducer gas. This was equipped for the determination of methane
and hydrogen by the explosion method.

An iron-constantan thermocouple was used in the main gas
flue for determining the temperature.

Two platinum platinum-rhodium thermocouples were used to
determine kiln temperature. One of these couples was placed
in each kiln door, 4' above the floor, and extended 6" into the
setting. The doors were closed by double 8" brick casings, with a
2" air space between.

An incline-tube draft gage was used for measuring the static
draft in the main draft flue.

A sampling device, similar to that used for the producer gas,
was employed for the flue gases.

An Orsat apparatus was used for the analysis of all of the flue
£as samples.

An iron-constantan thermocouple, for determining the flue gas
temperatures, was inserted in the draft flue beside the sampling
tube.

A Leeds and Northrup thermocouple-potentiometer, located
it a central station, was used in connection with all thermocouples.

(c) Procedure. — Atmospheric temperature and humidity read-
ings were taken every three hours.

Each charge of coal was weighed and a sample placed in a
sealed jar. A record of the coal charged during that period was
made each 6 hours. The samples of all coal charged during the
irst half of the test were combined into a single sample and
samples covering the second half combined into a second sample.
Proximate and ultimate analyses of these samples were made and
their heating values determined.

The crown of the main gas flue was uncovered about 3' west
Df the first branch flue leading to the kiln under test. This point
is designated (Fig. 3) as "Sampling Station A" and all tem-
peratures and samples of producer gas were taken at this location.

2500 cc. (cool) producer gas were drawn through the tar and
soot train every two hours and the temperature of this cooled
*as, so drawn into the aspirator bottle, was recorded.

The same procedure as above was followed in connection with
the moisture tube.

II

|< HJRNAL OF Till':

Analyses were made of producer gas samples covering periods
"i 6 hours, equal portions of the sample being drawn into thi
bottle ever) two horns. Analyses were made as soon us possible
after the samples were completed.

Temperature (potentiometer) readings were taken every 2
hours, covering the producer gas in the main flue; kiln (at both
doors), and the flue gases in the main elraft flue. Thermometer
readings at the cold junctions of all thermocouples were taken
at the same time.

The crown of the main draft flue was uncovered just outside
of the kiln hob, this point being designated (Fig. 3) as "Sampling
Station B," and all temperatures, draft readings and samples
of flue gases were taken at this location.

Draft gage readings were taken every two hours (Fig. 6).

7/

.19

II

/?

.16
\ IS

k a

1

—*/.

n ■;

v\ .07

"^ 06

\ .»

n

S

'

j

.02
.01

J\

1

1

Fig. 6. — Draft in the main draft flue.

Flue gas samples were taken in the same manner as the pro-
ducer gas samples and the analyses were made as promptly as
possible.

Barometric pressure readings were secured from the nearest
U. S. Weather Bureau Stations.

AMERICAN CERAMIC SOCIETY. 45

IV. The Test.

The handling of the burn was normal in all respects.

The kiln under test was burned for "reds" and contained
)522 smooth brick in the bottom, weighing 5.443 lbs. each, and
>2,i84 texture brick above, weighing 5.385 lbs. each.

The fires were started at 6 p.m. and burned for 95 V 4 hours.
L'he kiln which was being burned just prior to the test was cut
)ff at 2.30 p.m. and the producer remained idle until 6 p.m. Al-
hough instructions were given to maintain the fuel level in the
)roducer until the kiln to be tested was started, this was not done,
vith the result that the first six hours of the test showed a some-
vhat greater fuel consumption than would otherwise have been
lecessary and, of course, this extra fuel was charged against the
est. However, this was not a serious matter and does not
jreatly affect the figures.

The first settle measurement was taken after 71 hours and
howed 4". The further progress of the settle is indicated in
7ig- 9-

At 62 hours the static pressure of the gas at the burners was
).i4", water gage, and the air pressure, 0.95". At 72 hours the
>ressure of the gas in the gas main at "Sampling Station A" was

).I2".

CALCULATIONS.

1. Heat from the Coal Fired. — The total amount of coal
;upplied to the producer during the test was 32,106 kg.

During the first half of the test, 12,979 kg. of coal was used,
he average heating value being 6656 Cal. During the last
lalf 19,127 kg. was used, the average heating value being 6584

:ai.

2,979 X 6656 = 86,388,224 Cal.
9,127 X 6584 = 125,932,168

212,320,392 Cal. = total heat possible from the coal fired.

2. Volume of Dry Producer Gas.— The proximate analysis
>f the moisture free ash was

!'• JOURNAL OF THE

Per i I ni

Volatile matter i.xo

Fixed carbon 3.08

Ash 95 82

As the coal contained 4.58 per cent, ash

32,106 X 0.0458 = 1470 kg. pure ash in the coal.

Then — - — • = 1533 kg. ash from the coal supplied.
0.9582

1533 X 0.0308 = 47.2 kg. fixed carbon in the ash.

The main draft flue is said to fill up with tar and soot at the
rate of 12" in depth per year. The length of the main flue from
the producer to "Sampling Station A" was 130 ft. and the width
of the flue 5 ft. Therefore, the amount of tar and soot deposited
throughout this length of flue during the four day test can be
approximated as follows:

130X5 X — = 7.12CU. ft.
365

This material had a bulk specific gravity of 1.33.

. * . 7.12 X 62.4 X 1.33 = 591 lbs. or 268 kg. of tar and soot.

As about 93 per cent, of this is carbon, there occurred during
the test, a deposition between the producer and the sampling
station of 249 kg. of carbon.

As the total coal used was 32,106 kg.

32,106 X 0.6631 = 21,289 kg. of carbon in the coal fired.

21,289 — (47 + 249) = 2°»993 kg. of carbon passing the samp-
ling station in gaseous form and suspended in the gas as soot and
tar.

There was an average of 7.56 g. of tar in each cu. m. of pro-
ducer gas at the average flue temperature of 5800 C. This is
equivalent to 23.6 g. of tar in each cu. m. of producer gas at
o° C or 0.526 g. of tar in each gram-molecular volume of pro-
ducer gas at o° C. The results of a number of determinations
have shown that producer tar contains approximately 90 per cent,
carbon.

. ' . 0.526 X 0.90 = 0.4724 g. of carbon per gram-molecular
volume of producer gas at o° C.

AMERICAN CERAMIC SOCIETY.

47

There was an average of 3.47 g. of soot in each cu. m. of pro-
lueer gas at the average flue temperature of 580 ° C. This is
quivalent to 10.85 g. of soot in each cu. m. of producer gas at
>° C, or 0.242 g. of soot in each gram-molecular volume at

O f\

The total carbon from tar and soot per gram-molecular volume
if' producer gas at o° C was 0.4724 + 0.242 = 0.7144 g.

The average of the analyses of the producer gas2 samples
hows :

7.12 per cent CO2

2 1 . 85 per cent CO

3 . 25 per cent CH4

32.22 per cent, carbon-bearing gases.

Table i. — Analyses of Producer Gas.

Sample A.

C02 10.31

02 o . 80

CO 14 43

CH4 10.41

H2 8.93

N2 55 12

B.
IO.51
60
71
51
36
31

B.1

8.94

2.15

15 69

2-73

1538

55 ■ 1 1

C.

8.70

0.30

19.60

523

12 .70

53-47

D.

7 -5o

1 .60

17.40

4.90

1 1 .41

57 19

E.
7 30
0.30

19.20
683

10.44

55-93

Sample F. G

CO2 6 . 90 7 . 60

02 0.20 0 . 60

CO 24.30 20.10

CH4 8.70 8.60

H2 718 8.93

N2 52.72 54-17

H.
7 .60
3.30
3.30
3.36
7-53
7 9i

I.
7 .00
0.50
21.50

5-21

11.98
53-8i

J-

8.00

0.30

18.50

6.83

10.09

56.28

J-1

7.78

0.00

20.01

4.24

13-42

54-55

K.

6.90

0.20

19.70

8.30

8-79
56.11

Sample.
C02-.

02...
CO..

CH4.

H2...

N2...

L.

6.90

O.90

20. IO

8.17

8.65

55-28

M.
7 .00
o. IO

19.40
8.33
9.48

55 69

N.
IO
OO
80
64
OI
45

O.
7 . 20
0.20

16.90
9.85
6-53

59-29

O.1

7.04
0.23

2 I . 50

3-32
13.16
54-75

p

7.70

0.20

20.60

7- 15

8-33

56.02

1 Check analyses by the U. S. Bureau of Mines.

- Table I shows the analyses of the producer gas made at the time of the
est. Three of these samples were sent to the Bureau of Mines Laboratories
or exact analyees and these also are shown. As stated before, the field

48 JOURNAL OP THE

■

0.3222 X 12 = 3.8664 g. = weight of carbon in the carbo
bearing gases in one gram-molecular volume.

3.8664 + 0.7144 = 4.5808 g. total carbon in the producer
jjas per gram-molecular volume.

20,99^,000

1 hen • = 4,582,824 gram molecular volumes of drv

4.5808

producer gas,

4,582,824 X 22.3
or — = 102,197 cu. m. dry producer gas at o C,

1000

102,107 x

or — = ; — 7T-

273 273 + 580

x = 319,319 eu. m. dry producer gas at 5800 C, passing
"Sampling Station A" during the test.

3. Weight of Dry Producer Gas. — From the average analyses
of the producer gas

Per cent.
CO2 O.0712 X 44= 3.1328= 12. 41

O2 o . 0092 X 32 = o . 2944 = I . 1 7

CO 0.2185X28= 6.1180 = 24.24

CH4 O.O325 X 16 = O.52OO = 2 .06

H-2 o. 1365 X 2 = 0.2730 = 1 .08

X2 0.5321 X 28 = 14.8988 = 5904

25.2370 g. weight of
one gram-molec-
ular volume.

analyses were made with a Williams apparatus, using the explosion method
for determining hydrogen and methane. It is acknowledged that this method
is not reliable and a comparison of the analyses in this case shows that the
Williams apparatus gives high methane and low hydrogen in each instance.
Furthermore, this apparatus is not suitable for gases with high CO content
and further comparison of the analyses will show that the Williams apparatus
failed to indicate the full amount of CO present. In view of these errors,
it was decided to adjust the field analyses to correspond to the exact analyses
made in the laboratory, in the following manner: The field analyses of the
three samples, sent to the Bureau of Mines, were averaged, as were also the
analyses of the same samples made in the laboratory. The average of all the
field analyses was then adjusted in accordance with the relation between the
average of the three analyses made in the field and those made in the labora-
tory.

AMERICAN CKRAMIC SOCIETY! 49

4,582,824 X 25.237

1000
producer gas.

115,657 kg. = total weight of the dry

4. Free Hydrogen and Hydrogen in Methane in the Producer
Gas —

JI5>657 X 0.0108 = 1249. 1 kg. = free hydrogen.
115,657 X 0.0206 X \ 4 = 595.6 kg. = hydrogen in methane.

1844.7 kg. = total

5. Hydrogen in the Water Vapor of the Producer Gas. — There
was an average of 7.85 g. of water vapor accompanying each cu.
m. of dry producer gas at the average flue temperature of 580 ° C.

. " . 319,319 X 0.00785 = 2,506 kg. of water vapor in the pro-
ducer gas or 2,506 X Vg = 278-5 kg. = hydrogen in the water
vapor of the producer gas.

6. Hydrogen in the Water Vapor Entering the Producer with the
Air. — 4,582,824 X 0.5321 = 2,438,521 gram-molecular volumes
of nitrogen in the producer gas at o° C.

32,106 X 0.0152 = 488 kg. = nitrogen in the coal fired.

- = 17,425 gram-molecular volumes of nitrogen at o° C
28

from the coal.

2,438,521 — ■ 17,425 = 2,421,096 gram-molecular volumes of

nitrogen at o° C from the air.

= 3,056,939 gram-molecular volumes of drv air.

0.792

3.056.939 X 22.3

1,000

the producer.

68,170 x

— 68,170 cu. m. of dry air, at o° C, entering

273 273 + 24

x = 74,163 cu. m. dry air at atmospheric temperature (24 ° C ).

The vapor pressure of completely saturated atmosphere at
24 ° C is 22 mm. of mercury.

The average relative humidity during the test was 75.2 per
cent. (Fig. 7).

5°

JOURNAL OF THE

Tin' average partial pressure <>i" the atmosphere, due to the

moisture, was 22 X 0.752 = 16.54 nun. mercury.

The average barometric pressure during the test was 756 mm.
mercury.

756 — 16.54 ■ 739-46 nun.
756

74.163 X

240 C.

73946

= 75,822 cu. m. of moist atmosphere at

naye

•06/n

7? /??**

r

•ct/fy

ifi-^W

W

$

10
'2
it
17

'6

/' ^

W

./?/

f ; v

\fc

. i J

P

%

k*v

1 '

h\ j

j

$ or

n

y

\ *

#-"

^ €

\

f

'

It
II

i

S

;

Js •*"

f

*Ni2

Wj4

y

^2f2

22s

5^»

F^"^

1

I

t

1

1

X

O 6 /I /8 24 30 J* +2 44 S4 40 46 72 7» S4 SO S*

Fig. 7. — Temperature and Humidity of the Atmosphere.

During the test there was present an average of 16. 1 g. of water

vapor per cu. m. of atmosphere at 24 ° C.

. 75,822 X 16. 1 .

. . — — = 122 1 kg. of water vapor entering the pro-

1,000

ducer with the air, or

122 1 X = 135-7 kg. of hvdrogen.
9

7. Hydrogen in the Tar Carried in the Producer Gas. — If
the producer tar contained 90 per cent, carbon, its hydrogen
content was 10 per cent.

AMERICAN CERAMIC SOCIETY. 51

4,582,824 X 0.526 X 0.10 , , , , .

= 241 kg. of hydrogen in the tar.

1,000

8. Total Hydrogen in the Producer Gas. — The total amount
)f producer gas, passing "Sampling Station A" during the test,
contained the following hydrogen:

As-free hydrogen 1249. 1 kg.

In the methane 595 . 6 kg.

In the water vapor 278 . 5 kg.

In the tar carried in the gas 241 .0 kg.

2364.2 kg.

9. Hydrogen from the Steam Supplied to the Producer. — Of

;he tar and soot deposited in the gas flue during the test, about
j per cent, was hydrogen or, 268 X 0.07 = 19 kg. of hydrogen.

There was contained in the coal

32,106 X 0.061 = 1958.5 kg. hydrogen.

The hydrogen passing from the producer, other than that in
:he steam supplied for blowing was, therefore,

1958-5 — 19 + 135-7 = 2075.2 kg.

2364.2 — 2075.2 = 289 kg. hydrogen from the steam or, 289 X
} = 2601 kg. water, introduced into the producer, during the
lest, in the form of steam.

10. Heat Introduced into the Producer by the Steam. — As

:he pressure of the steam was 25 lbs. (gage) at the injector, the
leat carried in, in excess of sensible heat of water at atmos-
pheric temperature, was
2601 X (646 — 24) = 1,617,822 Calories.

11. Total Heat Introduced into the Producer. — 212,320,392 +
1,617,822 = 213,938,214 Cal. = total heat (above 240 C) in
:oal, steam, and atmospheric moisture, introduced into the pro-
ducer.

12. Heat Lost in the Ash. — From (2), there was 47.2 kg. of
ixed carbon in the ash drawn from the producer.

47.2 X 8100 = 382,320 Cal. in the ash.

382,320 ,

- — - — X 100 = 0.2 per cent. loss.

213,938,214

52 JOURNAL OF THE

13. Heat in the Dry Producer Gas. The total weight of dry
producer gas was 115,657 kg. The average temperature of the
producer gas was 580 ° C and, as the temperature of the atmos-
phere was 24 ° C, the mean temperature between the two was
302 ° C.

Sp. hi.

COj 0.1 ^4 1 X 0.244 = 0.0303

Oj 0.01 17 X 0.214 = 0.0025

CO 0.2424 X o. 256 = 0.0620

CH4 0.0206X0.834 = 0.0172

H2 0.0108 X 3.581 = 0.0387

N2 o . 5904 X o . 256 = 0.1513

0.3026 = specific heat of the dry pro-
ducer gas at the mean
temperature of 302 ° C.

IJ5)057 X (580 — 24) X 0.3020 = 19,420,198 Cal. = sensible
heat in the dry producer gas.

CO 115,657 X 0.2424 X 2,436 = 68,294,302 Cal.
CH4 115,657 X 0.0206 X 1 2,200 = 29,066,917 Cal.
H2 115,657 X 0.0108 X 29,100 = 36,348,682 Cal.

133,709,901 Cal. = heat from com-
bustibles in the
producer gas.

19,420,198 + 133,709,901 = 153,130,099 Cal. = total heat in
the dry producer gas.

14. Heat in the Tar of the Producer Gas. — 319,319 X 0.00756 =
2414 kg. = tar, carried past the "Sampling Station A," in the
producer gas.

Sp. ht.
2414 X (580 — 24) X 0.35 = 469,764 Cal. = sensible heat.
2414 X 8889 =21 458,046 Cal. = heat of combustion.

219,27,810 Cal. = total heat in the tar

15. Heat in the Soot of the Producer Gas. — 319,319 X
0.00347 = 1 108 kg. = soot carried past "Sampling Station A'
in the producer gas.

AMERICAN CERAMIC SOCIETY. 53

Sp. ht.

108 X (580 — 24) X 0.31 = 190,975 Cal. = sensible heat.

108 X 8056 = 8,926,048 Cal. = heat of combustion.

9,117,023 Cal. = total heat in the soot.

16. Heat in the Water Vapor in the Producer Gas. — From
5), there was 2506 kg. of water vapor in the producer gas, 122 1
:g. from the air entering the producer and the remainder (1285
:g.) from the steam and coal.

Figuring the sensible heat in 1221 kg. of water vapor above
4°C and the total (sensible and latent) heat in 1285 kg. water
rapor above 24 ° C (using the formula 0.447 ■*■ 324 X io~H for
pecific heat of water vapor)

221 X (580 — 24) X 0.545 = 362,917 Cal.
285 X 880 = 1,130,800 Cal.

i,493>7i7 Cal. = total heat in water vapor.

17. Total Heat in the Producer Gas at "Sampling Station A." —

[eat in dry gas 153,130,099 Cal.

leat in tar 2 1 ,927,810 Cal.

[eat in soot 9,1 17,023 Cal.

[eat in water vapor 1,493,717 Cal.

185,668,649 Cal. = total heat above 24° C.

18. Producer and Flue Losses. —

'rom (11), the total heat introduced into the pro-
ducer = 213,938,214 Cal.

rom (17), the total heat in the producer gas at
"Sampling Sta. A" = 185,668,649 Cal.

28,269,565 Cal.

X 100 = 13.2 per cent, loss (including the 0.2 per
213,938,214

ent. loss in the ash).

19. Heat Lost in the Flue Gases. — From (2), the amount
f carbon passing "Sampling ,Station A" in the producer gas was
o,993 kg.

54

JOURNAL OF THE

The average analysis of the flue gases (Table 2) shows:

Per 1 ' in .

CO 10.49 (by volunn-j

CO 0.12

CH< 0.10

10.71 of carbon-bearing gases.

Table 2. — Analyses of Flue Gas.

Sample

co2

o2

CO

CH4

H,

N2

Per cent, excess air

Sample

C02

o2

CO

CH4

H2

N2

Percent, excess air

4-i
16. 1

IS

79.8 80

c.
8.7
9 7
0.0

1 1

5 8

314.0 244.0
I

1.6

81.0

10

J.

5
6.7
0.0

82.8
44 o

K.

12 . 1

5-4
0.0

82.5
33 o

I).

4
7- 1
0.0

8l.5

49 o

h.

3
5-7
0.0

E.

1 1

5
6.4
0.0

82.1
42.0

F.
12.8

4-7
0.0

82.5
27 .0

N.

1 1

M.
1 1 .20 IO. O

4-0 3 7

83.O
35-0

0.2

o. 2

1 . 1

833

15-0

0.4

0.3
2 . 1

835
9.0

82
28.0

o.

12.5

0.0

0.6
0.5

3-2
83.2
-11. o -

H.

13-6

3-6

O.O

82.8
20.0

P.

9.2
O.O
0.8
0.6

4-2

85.2

-14.0

0.107 1 X 12 = 1.2852 g. = weight of carbon in the carbon-
bearing gases in one gram-molecular volume of flue gases at
o°C.

— = 16,334,423 gram-molecular volumes of dry flue

1.2852

gas.

The weight of one gram- molecular volume of dry flue gas

was

Per cent.
4.6156 = 15.500
I .9776 = 6.650
O.O336 = O. 112
O.O160 = O.054
O.OI32 = O.O44

C02 o. 1049 X 44 =

02 0.0618 X 32 =

C02 0.0012 X 28 =

CH4 0.0010 X 16 =

H2 o . 0066 X 2 =

N2 0.8245 X 28 = 23.0860 = 77 .640

Total, 29.7420 g. 100.00

AMERICAN CERAMIC SOCIETY. 55

.'. 16,334,423 X 29,742 = 485,818 kg. = the total weight of dry
ue gases. As the average temperature of the dry flue gases was
44.7 ° C, and of the atmosphere 240 C the specific heat of the
ases at the mean temperature (1840 C) was

Sp. ht.

O2 o. 15500 x 0.2239 — 0.0347

i2 0.06650 X 0.2200 = 0.0146

O o. 001 12 X 0.2507 = 0.0003

H4 0.00054 X 0.7456 = 0.0004

[2 o . 00044 X 3 . 5 104 = o . 001 5

rj 0.77640 x 0.2507 = o. 1946

0.2461 = specific heat of the dry gases.

485,818 X (344.7 — 24) X 0.2461 = 38,342,831 Cal. = sensible
teat in the dry flue gases.

O 485,818 X0.00112 X 2,436 = 1,325,467 Cal. in the uncon-

sumed CO.

H4 485,818 X 0.00054 X 12,200 = 3,200,569 Cal. in the uncon-

sumed CH4.

[2 485 ,818 X 0.00044 X 29,100 = 6,220,414 Cal. in the uncon-

sumed H2.

10,746,450 = Cal. heat in the un-
consumed com-
bustibles in the
flue gases.

16,334,423 X 0.8245 = 13,467,732 gram-molecular volumes of
dtrogen in the flue gases.

Deducting the nitrogen in the producer gas as found under
6).

13,467,732 — 2,420,398 = 11,047,334 gram-molecular volumes
f nitrogen.

= 13,948,654 gram-molecular volumes of dry air,
0.792

tihich entered the kiln through the burners.

13,948,654 X 22.3 . 0

— - = 311,055 cu. m. of dry air at o C.
1000

3H,055 = *

273 273 — 24

56 JOURNAL OV THE

v - 338, |<»> CU. in. dry air at 240 C.

7 56
538,400 X •- = 345,969 cu. m. of moist atmosphere at 24 CS

73946
As there was present an average of 16.1 g. of water vapor per

c * t ,f i 345»969 X 16.1 . ,

cu. m. ol atmosphere (Fig. 7), = 5570 kg. of

1,000
water vapor entering the kiln with the air required for com-
bustion and the excess air.

From (8), the total hydrogen in the producer gas was 2364.2

kg-

The hydrogen, other than that in the moisture, escaping the
flue gases was

CH4 485,818 x 0.00054 x Vie = 65.6 kg.

H2 485,818 X 0.00044 = 213 .6 kg.

279.2 kg.

Assuming that all of the free hydrogen found in the flue gases
came from the producer:

2364.2 — 279.2 — 2085 kg. of hydrogen or 16,680 kg. of water
vapor in the flue gases coming from the producer gas. 5570 +
16,680 = 22,250 kg. of water vapor, which left the kiln in the
flue gases in addition to that which came from the ware.

The moisture entering with the air at the producer and with the
air for combustion at the kiln was 1221 + 5570 = 6791 kg.

22,250 — 6791 = 15,459 kg. of water vapor evaporated or
produced within the producer and the kiln.

L 6791 X (344-7 — 24) X 0.507 = 1,104,272 Cal. sensible heat
above 24 ° C in the water vapor from the air.

J5)459 X 740 = 11,439,660 Cal. total heat above 240 C in the
water vapor from the steam and coal.

38,342 ,831 Cal.

10,746 ,450 Cal.

1,104 »272 Cal.

11.439 »66o Cal.

61,633,213 Cal. = total heat ost in the flue gases other than that
leaving in the water vapor from the ware.

AMERICAN CERAMIC SOCIETY.

57

61,633,213

X 100 = 28.8 per cent. loss.

213.938,214

20. Heat in the Mechanical and Hygroscopic Water. — The

:iln setting consisted of 71,706 brick, which weighed when burned
75,402 kg.
The ignition loss of the clay was 3.39 per cent.

.". -^^ = 181,557 kg. = weight of the green ware set

1. 00 — 0.0339

dry basis).

The green ware as set contained 0.543 Per cent, mechanical and

ygroscopic water.

.'. ; — = 182,548 kg. = weight of the damp green

1. 00 — - 0.00543

rare as set.

182,548 — 181,557 = 991 kg. = mechanical and hygroscopic
rater in the ware.

Assuming the mean temperature of evaporation of the me-
hianical and hygroscopic water to be 1250 C (the corresponding

Fig. 8.— Producer gas and flue gas temperatures and excess air.

58 l« lURNAL OF THE

temperature of the flue gases (Fig. 8) being about 60 ° C), the heal

lea\ ing the kiln in tliis water vapor was

991 x (u.s 24) = 100,091 Cal.
991 X 522 = 517,302 Cal

6l 7.393 Cal.

As this water vapor cooled down to 60 ° C, it gave up heat as
follows:

991 X (125 — 60) X 0.477 — 28,812 Cal.

.'. 617,393 — 28,812 = 588,581 Cal. = net loss.

588,581 X 100

213,936,214

0.3 per cent.

21. Heat Consumed in the Dehydration of the Clay and
Leaving in the Vaporized Chemical Water in the Flue Gas. —

From (20) the weight of the dry green ware was 181,557 kg.
The burned weight of the ware was 175,402 kg.

.'. 181,557 — 175,402 = 6155 kg.. = approximate weight of the
chemical water.

The temperature of the maximum rate of dehydration of clays
is about 560 ° C.

6,i55 X (560 — 24) X 0.2 = 659,816 Cal.
6,155 X 609 = 3,748,395 Cal.

6,155 X 261 = 1,606,455 Cal.

6,014,666 Cal.

The flue temperature corresponding to a kiln temperature of
560 ° C was (from curve — -Fig. 8) about 1500 C and the heat
given up by this water vapor, on cooling from 5600 C to 1500 C,
was

6,155 X (560— 150) X 0.562 = 1,418,235.

.'. 6,014,666 — 1,418,235 = 4,596,431 Cal. = net consumption.

4.596,43! X 100

213,938,214

= 2.2 per cent.

22. Heat in the Ware at the End of the Burn. — At the end

of the burn the temperature of the ware 4' above the floor

AMERICAN CERAMIC SOCIETY.

59

veraged 1 135 ° C (Fig. 9), the temperature through the center
i the crown was 1 135 ° C, and the temperature at the floor level
ras 1070 ° C. Assuming that the level of 4' above the floor
ivided the mass of ware into halves, taking the temperature

Ce/r/e

r ^ '&

*'M

^ZZ*

<?*//&<:

u

•m

700

4

soc

"a

y

/

V

J>

s

1/

'

1

t

Y

/oo

I

1

0

X

Fig. 9. — Internal Kiln Temperatures and Settle.

f the upper half as 11350 C and the temperature of the lower
alf as the average between 11350 C and 10700 C or approxi-
lately 11000 C, then the temperature of the whole mass can be
onsidered as the average between 11000 C and 11350 C or
ii8°C.

As the weight of the burned ware was 175,402 kg., the heat
^presented in this ware at the end of the burn was

175,402 X (11 18 — 24) X 0.23 = 44,135,011 Cal.

44,135,011 X 100 , ,

^ ^° ■ = 20.6 per cent.

213,938,214

23. Heat in the Kiln Structure and in the Ground at the End
f the Burn and the Radiation Loss. — The losses previously
alculated amount to 65.1 per cent.

100 — 65.1 = 34.9 per cent, of the total heat was stored up at
tie end of the burn in the kiln structure and in the ground under
nd around the kiln and was lost by radiation during the burn.

Go

rOURNAL OF THE

As soon ;is the fires wirt' cut off :il (lit end of the burn, t< m

perature readings of the external surface of the kiln wen- taken at

numerous places on the crown and side walls, by means od a

copper constantan couple, of small gauge, and a potentiometer;

The average temperature of the exterior of the crown was

i ;<> C , the exterior of the side wall above the hob 78 ° C, and
the exterior of the hob 46 ° C. Knowing the internal and ex-
ternal temperatures, curves were drawn indicating the probable
temperature throughout the cross section of the several parts
and, from these, average mass temperatures were secured. Crown
4770 C, side wall 42c;0 C, and hob 423 ° C. Then, the heal
contained in the kiln structure above the ground was as fol-
lows:

Crown 49.432 X (477 — 24) X0.23 = 5,150,320 Cal.

Side wall 45 ,562 X (429 — 24,) X 0.23 = 4,244,100 Cal.

Hob 45 ,017 X (423 — 24J X 0.23 = 4,131,210 Cal.

13,525.630 Cal.

13,525,630 X IOO

= 6.3 per cent.

213,938,214

The heat stored up in the kiln bottom and in the ground,
below and around the kiln, and the heat radiated during the burn
was

34.9 — 6.3 = 28.5 per cent.

/00%

•A

/J.2> fZvabcer <?/7<y ' /^/t/e /ass

2$s\ loss //? /^/t/e teases

] /Le0t/s/7& A//? //7 l/a/7c>r/?ea/ /Wec^a/p/co/ o/7</ /faar0scc>£>/c

\ (?0/7Jt//77e& ' //? y?e/yc/sv//c>/7 a/ C/ay <?/?</ /eoyz/Tp //7
z'2) J4?/?£?sy/ee/ cAe/77/ca/ State/- /h Syfae &ajes

2iu\ -ft? Jtas-& a/ '/ypa1 a/ 'j&t/sv7

6 J > //7 A/'//? ^/y&cTc/s-e rfApfe /y?e £r/~w/?a' ' a/'/vra'qf^&s-/?

) f/7 /9/V/7 £o/%7/77 a/7l///7 tf/W/?*/ 0/*/r/7a' of/fff/? &/?</

J /oss <6y A^a'/£//Pa/p a/7t/ &<?/7<yec//o/7 <?t/r//?y J%/r/7
Fig. 10. — Heat Distribution.

AMERICAN CERAMIC SOCIETY. 61

V. Conclusion.

The coal consumption was 989 lbs. per 1000 brick.

No direct means was available for measuring the steam sup-
lied to the producer, hence it was necessary to derive the amount
y calculation. The results indicate that a smaller quantity of
:eam was used than in normal producer practice. This was to be
cpeeted, however, as the steam was cut off from the bosh and
itered only through the center tuyere.

Using the external kiln temperatures, secured at the end of the
urner, time-temperature curves were drawn indicating the
robable surface temperatures throughout the burn for the several
arts of the kiln. From these curves were secured the average
?mperatures as follows: Crown 74 ° C, side walls 41. 6° C, and
ob 31 ° C. Using the formula of Peclet for radiation and con-
action, the total radiation and the convection loss from the
iterior surface of the kiln was calculated to be 4.8 per cent,
his is undoubtedly too low and the figures were not included
> a part of the calculation. However, attention should be called
) the unusually low external kiln surface temperatures, which,
)gether with the short duration of the burn, would necessarily
'suit in a low radiation loss. In the author's opinion, the total
idiation and convection losses from the kiln structure, in-
.uding the burners, was not over 15 per cent, and possibly as
>\v as 10 per cent.

The coal was elevated mechanically to the producer charging
x>r and all coal and ash handling was centralized.

The advantage of a clear kiln yard, with entire absence of coal
id ash piles, need hardly be argued.

Where no space is necessary for piling coal, the kilns may be
aced closer together, resulting in a more compact kiln yard.

As the producer and flue loss, as indicated in the calculation,
as over 13 per cent., the use of a producer, from the standpoint
' heat generation only, cannot be considered more efficient than
ie combustion of coal in the kiln furnaces, but an advantage
?s in the ability to control the application of the gas to the
his. This is brought out clearly by the record of excess air in
ie flue gases, indicated in Table 2 and Fig. 8. This air excess
much lower than in most direct-coal-fired kilns.

JOURNAL OF THE

Either oxidizing or reducing conditions, to any degree, may
be had at will and change from the one condition to another can
be effected almost instantaneously.

Only a limited amount of solid fuel can be burned in direct
lircd-kiln furnaces. Except by increasing the draft through tin
grates, the amount of fuel consumed can be increased only by
increasing the grate area. In the gas-fired kiln, there is no
such limitation, with the result that the heat application, at safe
periods during the burn, can be made at a much greater rate than
in the case of solid fuel direct-fired.

One burner and one producer man handled the burning in 12-
hour shifts. The boiler fireman removed the ashes from the
producer. On Sundays, the day burner looked after the boiler
and removed the ashes from the producer, if necessary.

When the number of direct-fired kilns is increased to such ar
extent that one burner cannot take care of them, then the pro-
ducer-gas-burning system would show a saving in labor, but
there is no great relative reduction in labor for a small gas in-
stallation.

Producer gas has a lower flame temperature than coal, direct
fired. Producer gas with its inert constituents — nitrogen, carbon-
dioxide and water vapor — loses a considerable part of its sensible
heat in traveling from the producer to the burners, in addition to
the loss due to "cracking." These inert gases, acting as a dilutant
to the combustible mixture, reduce the resulting combustion
temperature. In view of this, it is unreasonable to expect very
high temperatures from producer gas, when using air at atmos-
pheric temperatures to support combustion.

The advantages and disadvantages of this system may be
summarized as follows:

Advantages:

1. Centralization of coal and ash.

2. Clean kiln yard.

3. Reduced kiln yard space.

4. Combustion control.

5. Generation of a large amount of heat in a limited furnace
space.

AMERICAN CERAMIC SOCIETY. 63

6. Quick burns.

7. Labor saving when the capacity exceeds a one-man unit.

Disadvantages :

1. Producer and flue losses.

2. Lower flame temperature.

DISCUSSION.

Mr. Blair: I would like to ask Mr. Harrop if his producer
)sses were offset by the better control of excess air in the kiln?
rou stated that the grate would burn with approximately 900
ounds of coal per 1000 brick. What could they do with coal
ombusted in the fire boxes direct? Could they do as well as
lat?

Mr. Harrop: In my judgment, the saving due to the well-
igh perfect control that they were able to get in the way of
xcess air and the ease of manipulation of the combustion condi-
ions would fully cover the producer losses. I am sure that if
he saving in the reduction of excess air would not cover the
roducer losses, that the other advantages would. It could be
alculated, quite readily, whether the expenditure of fuel to heat

normal amount of excess air over what we found, would equal
r be in excess of 13 per cent. It would be very interesting
d see whether it would cover or exceed the loss in the producer.

Mr. Montgomery: I would like to ask Mr. Harrop to please
epeat his statement in regard to the maximum temperatures
eached in a periodic kiln, using producer gas with cold air. He
aid something about reaching cone ten in the bottom of the
iln.

Mr. Harrop: That was a continuous kiln which was being
red with a very poor grade of lignite, and the man who reached
one ten said he stopped the kiln, not because he could not reach

higher cone, but because he feared the kiln would be damaged.

Mr. Minton: I would like to ask, what was the height and
iameter of the stack and how high were the inside walls?

Mr. Harrop: The stack is 40 feet high and I think the
ivision walls extend to the top of the stack.

64 AMERICAN CERAMIC SOCIETY.

Mr. Ortman: It seems to me thai ;i statement from Mr.
I l;n i < >]) as to his observations of the temperatures of the firJ

boxes in tliis case as compared with observations made on dii
coal fired methods and the probable effect that the differeno m
temperature, if any, would have on kiln repair, might form I
very valuable addition to the record in connection with tins
kiln.

Mr. Harrop: This test was made about a year and a hall
ago — my remembrance is that there were no enclosed fire boxes,
similar to those used for burning solid fuel, but that the hole or
opening through the side wall was very small. There was a bag
wall, to be sure, but I do not believe that the temperatures reached
in what you might call the lire boxes are as high as the tem-
peratures reached in the combustion of solid fuel.

NOTES ON THE HYDRATION OF ANHYDRITE AND
DEAD-BURNED GYPSUM.

By A C. C.ii.l, Ithaca, N. Y.

In the prosecution of certain studies which, in accordance
dth the plans of the late H. E. Kramm, of Cornell University,
rere to have led up to an exhaustive discussion of the proper-
ies and conditions of the formation of gypsum and anhydrite,
wo experiments were inaugurated, the results of which have
ecently come to light. These seem worthy of record both on
.ccount of their intrinsic interest and as a tribute to the pains-
aking methods of Mr. Kramm. The object of this investiga-
ion seems to have been to determine the extent to which time
light be a factor in the "setting" of anhydrite and dead-burned
ypsum. As is well known, for short periods of time, these sub-
tances show little or no tendency to set.

The experiments were conducted in two wide-mouthed bottles
>r cylindrical jars having metallic screw-caps, but without rubber
ings for air-tight sealing. Each jar was a little more than one
nd a half inches in diameter, and slightly less than six inches
ligh, having thus a capacity of about 250 cc. In one was placed
ome 50 or 60 grams of anhydrite from Windsor, N. S., broken
ipproximately fine enough to pass through a sieve of six or eight
neshes to the inch. The other was charged with several lumps
)f Hillsboro, N. B., alabaster which had been heated for three
ind a half hours to a maximum temperature of 560 ° C, the lumps
>eing an inch or more in greatest diameter, and having an aggre-
;ate volume of perhaps 60 or 75 cc.

Both bottles were filled with water and closed on Oct. 28,
911. Uate in 1917, a little more than six years after the begui-
ling of the experiment, they were discovered with their carefully
>repared labels still intact. A striking appearance was pre-
ented, as shown by the two photographs (Figs. 1 and 2) which
vere made after cutting off the tops of the bottles. The water

66

J01 RNAL OF THE

had entirely evaporated from both specimens. They may
perhaps be best described separately.

I. Anhydrite. (See Fig. i.)
The small fragments of anhydrite were so firmly bound to-j
gether that it was impossible to pry them out for analysis without
breaking to a coarse powder. Glistening crystals of gypsum
could be seen in the interstices, and brilliant, acicular crystals,
attaining a maximum length of about one-half inch, projected

Fig. i.
Consolidated anhydrite with gypsum needles, formed
within a period of six years.

into the free space which had been occupied by the water . These
crystals gave very good reflections on a Fuess goniometer, showing
by their prism angle of 68° 39/, that they are elongated in the
direction of the c axis.

In order to determine to what extent hydration had taken
place, about half a gram (0.4475 g-) °f the material was dug out
with a heavy knife point, avoiding as" far as possible the visible

AMERICAN CERAMIC SOCIETY. 67

gypsum crystals. On heating to constant weight at low red
heat the loss was found to be 8.29 per cent. (37. 1 mg.). This
:orresponds to a hydration into gypsum of 34 . 15 per cent, of the
Driginal anhydrite. It shows conclusively that the hydration
tvas not limited to the amount of calcium sulphate which could
ae held in solution at one time, since all the water originally pres-
ent would dissolve less than half a gram of that substance, while
ifteen or twenty grams of the anhydrite must have been hy-
irated.

Unfortunately, no data is at hand to indicate howr much of
the "setting" was due to temperature changes, thus leaving in
loubt the extent and rate of recrystallization which would take
jlace at constant temperature. That the anhydrite fragments
^ere firmly cemented by gypsum crystals, under the existing
conditions, is fully established.

II. Dead-burned Gypsum. (See Fig. 2.)

The pieces of burned alabaster in the second bottle had become
veritable porcupines with a complete covering of fine bristling
jypsum needles about one-eighth of an inch in length. Inside
:he spiny coating, the mass had "set" to much greater hardness
han ordinary plaster of Paris. By means of heavy forceps a
>ortion was cut from one of the lumps, and, after the crystals
vere broken away from its surface, a water determination was
uade by the method of loss. Nearly a gram of the material
0.9865 g.) lost 18.8 per cent. (185. 4 mg.). This indicates a
e-hydration of 87.46 per cent, of the calcium sulphate of the
turned gypsum. The difference in behavior between this sub-
tance and the anhydrite is very marked, the water seemingly
mding its way back to its place in the dehydrated gypsum much
more readily than it could make place for itself in the crystallized
nhydrite. The smaller amount of water, larger amount of cal-
ium sulphate, and the greater degree of hydration, make the
rgument against simple solution and subsequent evaporation
ven more compelling in this case than in that of anhydrite.

A further phenomenon should be noted as having a bearing
n the significance of this experiment. The bottle was found to
ave a vertical crack from the top almost to the bottom. Through

68 JOURNAL OF THE

it the water had evidently oozed and by evaporation a small
amount of gypsum was deposited on the outside of the bottle.
It seems almost certain that this crack was produced by the ex-

Fig. 2.

Fragments of "dead-burned" gypsum, with coating

of reerystallized gypsum needles.

pansive force of hydration or by the growing force of the gypsuni
crystals. Mr. Kramm would hardly have used a cracked bottle
and if the crack had existed during the whole of the experiment,

AMERICAN CERAMIC SOCIETY. 69

it would seem that the water would have escaped before pro-
ducing such extensive changes. The crack would allow of more
rapid desiccation, however, and this may explain the smaller
size of the gypsum needles in this bottle.

The fact that, in both bottles the multitude of acicular crys-
tals were all attached to the masses of introduced sulphate and
none to the glass walls, would seem to corroborate Taber's view —
that the crystals grow at the point of attachment rather than at
the free end.

In conclusion, while emphasizing the fact that under the condi-
tions of these experiments both natural anhydrite and "dead-
burned" gypsum attained a firm "set," it may be pointed out that
there is room for further investigation as to the rate at which this
takes place, since no observation was made as to the time when
evaporation was complete. Moreover, the question of the rela-
tive importance of temperature changes, causing supersatura-
tion by cooling, and of constant temperature hydration and crys-
tallization, due to the greater solubility of CaS04 than of CaS04 +
2 HoO, is raised rather than answered.

COMMUNICATED DISCUSSIONS.

F. A. Kirkpatrick: The hydration of anhydrite is interesting
from a geological point of view. Certain gypsum deposits have
evidently been formed by the weathering of the anhydrite which
is found beneath the gypsum. The rate of hydration in nature
is probably much slower than that observed in the laboratory, as
the conditions in the latter case were very favorable for hydra-
tion. Anhydrite is not used as a material of construction since
it hydrates and sets too slowly.

The hydration of burned gypsum presents a practical problem
of great interest. There still remains a large amount of experi-
mental work to be done along this line. The well-known in-
vestigation of Glasenapp on Hydraulic Gypsum should be ex-
tended to include strength tests. If we should accept Glase-
napp's classification of burned gypsum products, the sample of
Hillsboro alabaster heated for three and a half hours to a max-
imum temperature of 560 ° C would be completely dehydrated.

70 JOURNAL OF THE

In contact with water it would show, in short periods of time, no
hardening or only imperfecl hardening. It would contain little
or no hydraulic gypsum. In the long period of storage, about
six years, the material used would hydrate and set slowly, proba-
bly mostly by crystallization. The introduction of alum or other
catalyzers would decrease the time of set sufficiently to allow the
material to be used for practical purposes.

W. E. Kmi.hy: Mr. Gill has established one point which is
of great interest and importance to the gypsum industry, viz.,
! that both anhydrite and dead-burned gypsum will hydrate if
given sufficient time.

The ordinarily accepted theory of the setting of gypsum, as ap-
plied in this case, would be that the water dissolves the anhy-
drous calcium sulphate which, subsequently, crystallizes out in
the form of the less soluble hydrated gypsum.

Theoretically, a small amount of water could continue to dis-
solve anhydrite and deposit gypsum indefinitely. The quantity
of water required is not at all dependent upon the solubility of
the anhydrite, but depends solely upon the fact that the gypsum
which is deposited carries with it two molecules of water for each
molecule of calcium sulphate.

Mr. Gill leads one to infer that he believes his experiments
have thrown some doubt upon the working of this theory; I
cannot see that this is the case. I believe that the fact that his
dead-burned gypsum hydrated to a greater extent than his an-
hydrite is due to the greater speed of reaction rather than to the
relative quantities of water present. I believe that in both
cases, the reactions continued until the water left as such was not
sufficient to supply the two molecules necessary for the crystal-
lization of the calcium sulphate held in solution.

The dead-turned gypsum was apparently more active than the
anhydrite, and therefore was able to hydrate to a greater extent
before the water had evaporated.

I believe that the experiment does prove conclusively that
anhydrite and dead-burned gypsum will set; I believe that it
also proves that dead-burned gypsum, formed under the condi-
tions of the experiment, sets more rapidly than anhydrite.

AMERICAN CERAMIC SOCIETY. 71

I do not believe that the experiment throws any particular
light on the mechanics of the setting, because the two most im-
portant elements were not observed, viz., the time and the tem-
Derature.

A. C. Gill: The discussions of Messrs. Kirkpatrick and
Emley lead me to emphasi2e tte fact that the experiments with
anhydrite and dead-burned gypsum were inaugurated and planned
jy Mr. Kramm. The present paper is merely an attempt to
salvage some of the results which he would have attained if he
lad lived to complete the investigation.

Mr. Kramm doubtless intended the time and temperature of
heating the Hillsboro alabaster to insure complete dehydration.
He may even have carried out analyses to confirm this, though
10 such record is known to me. The possibility of using alum
is an accelerator, as suggested by Mr. Kirkpatrick, is worthy of
:areful consideration.

As to Mr. Emley's discussion, it may be pointed out that the
formation of gypsum crystals, to which the setting of the substances
mder investigation is supposed to be due, may take place from
solution or simply by the addition of two molecules of water to
;ach one of calcium sulphate, without the latter substance pass-
ng into solution. If from solution, it doubtless would go on at
:onstant temperature when saturation for calcium sulphate
NB.S reached. But what the relative rate of setting would be at
:onstant temperature, as compared with that produced by the
liurnal temperature change of the laboratory, does not appear
rom the experiment. Nor is there any indication of the relative
mportance of setting without solution. Under the conditions
)f the experiment, it is certain that the relative quantities of
vater present may have had a great influence on the time re-
mired to reach saturation.

JOURNAL OF THE

MEETINGS OF THE LOCAL SECTIONS,
AMERICAN CERAMIC SOCIETY.

(Pull accounts of all meetings should he- sent to the Bditor, Bos 444,
New Brunswick, N. J.)

NORTHERN OHIO SECTION.

A meeting of the Northern Ohio Section was held in the afterno
evening of June loth, in Cleveland, Ohio. During the afternoon the fol-
lowing plants were visited:

Cleveland Metal Products Company, manufacturers of enameled lamj
shades, stove parts, etc. The operations include pressing, oxyacetyleni
welding, electrical welding and enameling.

National Lamp Works, Euclid Glass Division, manufacturers of glass
electric light bulbs. The points of interest were a Smith gas producer plant
the molding of glass melting pots and the manipulations in the manufacture
of glass tubing and glass bulbs.

Nela Park Auditorium, Nela Park (Cleveland research laboratories of the
General Electric Co.). An interesting illustrated lecture on" the manu-
facture of electric lights was delivered by a member of the General Kleetric
staff.

A joint meeting of the Section and the Cleveland Section of the America?
Chemical Society was held in the evening at the Olmstead Hotel. An in-
structive address on "Petrographic Studies in Ceramics" was delivered by
A. A. Klein of the Research Laboratories, Norton Company, Worcester
Mass.

B. A. Rice, Secretary.

JOURNAL

OF THE

\MERICAN CERAMIC SOCIETY

A monthly journal devoted to the arts and sciences related to
ne silicate industries.

bl. 1 February, 1918 No. 2

EDITORIALS.

UNITY AND ECONOMIC LEADERSHIP FOR THE
CERAMIC INDUSTRIES.

The war has brought out rather strikingly the curious lack of
nity of purpose among the several branches of our industry,
^en among those manufacturing the same general type of prod-
ct. The absence of united action on the large economic ques-
ons confronting us to-day has been very apparent, especially
i dealing with the several Governmental bodies which now con-
■ol our industrial destiny. Even though some of our industries
ave organized more or less effectively, these associations have
een lacking in cohesion and harmonious cooperation with al-
ed ceramic branches. The result has been the failure of this
nportant group of industries to secure representation on the
/ar Industries Board.

It was only at the instance of a member of the Industrial
oard, himself not familiar with the clay industries, that manu-
icturers of structural ceramic products were brought to the
salization of the necessity for united action. This affords a
;riking lesson showing the need of a more general association of
iterests. In the American Ceramic Society we have already
presented the technical interests of all the ceramic branches,
ut we must have an avenue also for the discussion of broad
xmomic subjects and for concerted action when the occasion
-ises. This might be accomplished through the agency of this
Dciety, perhaps through an Economic Section or through an

7i JOURNAL OF THE

inter-organization of all the manufacturers' associations, a unioi

ceramique.

This brings up also the question of leadership. It is evident
that the greatest amount of good can be accomplished only through
the medium of broadly trained men for the positions of chair-
men and secretaries of the different subdivisions, men who com-
bine a thorough knowledge of business and markets with sound
information concerning the general aspects of the important
technical questions, such as those relating to fuel economy, labor-
saving machinery, and last, but not least, concerning the great
problem of labor. Such men are rare, and if they are not always
available they must be trained for such work as far as is possi-
ble, by cooperation between the manufacturers and the col-
leges. It is evident that such individuals, to be useful, must
not be clerks or salesmen, per se, but must be the resultant of a
combination of those innate qualities which mark men of broad,
human vision, and energy. Without such men, and we already
have some in the industry, organized effort will prove to be
a mere empty shell. Ex nihilo nihil fit.

ACID RESISTING ENAMEL WARE.

The production of acid-resisting enamel-lined equipment in
the United States has been subject to very definite wartime in-
fluences. These influences were felt soon after the outbreak of
the war in August, 19 14. Up to that time greater progress had
been made in the manufacture of enameled steel apparatus for
the various conditions of service ranging from the mild condi-
tions encountered in the preparation of food to the severe mineral
acid conditions encountered in the manufacture of gases, etc.
vSome cast iron equipment, lined with a relatively soft enamel,
was produced at that time in this country, but most of the enam-
eled cast iron wares used under the more severe chemical condi-
tions were imported from Austria, Germany and France, particu-
larly the latter country.

Shortly after the outbreak of the wrar, the manufacturers of
acid-resisting enamel ware were confronted with two condi-
tions: First, with the development of the dye and explosive in-
dustries there was a greatly increasing demand for acid-resisting

AMERICAN CERAMIC SOCIETY. 75

lamclware equipment. vSecond, the supply of the imported
ares was greatly reduced, and the failure to secure a supply
reatly inconvenienced some manufacturers who had standard-
;ed their processes with a view to the continued use of the cast
on apparatus of foreign design and manufacture. These con-
itions stimulated the production of domestic equipment and a
larked improvement in the quality resulted but substitutes
>r the cast iron imported wares were not at first placed upon the
larket.

Owing to the heavy demands for steel plates in the munitions
idustries and later in the shipbuilding industry, a decrease in
le quality of the plates available for the manufacture of enameled
are was noted for a time. Simultaneously with the decrease
1 quality of the plates the demand for high quality enameled ap-
iratus for use in severe chemical service greatly increased.
here appeared on the market at this time a small quantity of
merican made cast iron apparatus so constructed and enameled
> to excel the wares formerly imported. The users of this type

apparatus quickly realized the merits of the cast iron ware
anufactured in this country7 and the demand increased rapidly.

is at present available in units of quite large capacity, up to
)o gallons, and the limit has not been reached.

Since the outbreak of the war, an increased field of usefulness
r enameled steel apparatus has been in the food industries
id it is now used in the processing of materials whose mild corro-
ve action permits the use of relatively soft enamels. With the
tpid expansion of the dairy industry and the increased manu-
.cture of butter substitutes, including the hydrogenation of
Is, large enameled steel units up to 5000 gallons capacity are
; great demand.

It is gratifying to note that the acid-resisting enamel equip-
ent industry is solving its problems in this time of effort. Manu-
.cturers of pure food, explosives, gas for warfare purposes,
larmaceutical preparations, dyestuffs and of many other
les, find in the acid-resisting enameled iron wares a very useful
id essential type of equipment.

The future of the acid-resisting cast iron branch of the enamel-
g industry in this country is most promising.

ORIGINAL PAPERS AND DISCUSSIONS.

ON THE RELATION BETWEEN THE PHYSICAL PROPER-
TIES AND CHEMICAL COMPOSITION OF GLASS, VIII:
MOLECULAR COMPOUNDS.

By Edwin Ward Tillotson, Jr.

This paper is an account of an attempt to secure information
on the chemical compounds which arc formed in glasses. The
fundamental law which makes this possible has been extensively
tested with mixtures of liquids and solutions and, without going
into a mathematical demonstration, it may be said that theo-
retical considerations suggest that the refractive index of a mix-
ture should be a linear function of the relative volumes of the
components of the mixture, and this prediction is verified by the
experimental data for a large variety of mixtures. When the
refractive indices and the composition of a series of binary mix-
tures are plotted in a suitable manner, one or more straight lines
are usually formed, the intersections of which, if there be more
than one straight line, are interpreted as indicating the formation
and the composition of the molecular compounds which consti-
tute the mixture. This relationship, therefore, offers a means
for obtaining information regarding the chemical condition of
mixtures from which, because of the physical condition of the
mixture, or of their own specific nature, the compounds formed
therein may not be separated.

One of the most important applications of this law is i r tl
study of glasses. Heretofore there has been no means by which
the compounds present in a glass could be identified. Such in-
vestigations as have been carried out have been along two dis-
tinct lines: first, the behavior of glass surfaces toward reagents:

AMERICAN CERAMIC SOCIETY. 77

i second, the melting-point diagram of binary mixtures of
nparatively simple and crystallizable silicates, borates, and
Dsphates.

fhe result of the former line of investigation has been the
)ption of the following formula as typical and ideal for a nor-
1 glass: Na2O.Ca0.6Si02.1 Although it may have been origi-
ly intended that this should represent merely the approximate
nposition of a good glass, the frequent use of this formula in
;mical literature gives one the impression that a good glass
mid consist of a single chemical compound, and that a glass
>sessing this formula is such a compound, and therefore a
)ical and ideal glass. A careful examination of the data on
ich this formula is founded, fails, however, to reveal any con-
.cing evidence that it represents a definite chemical com-
ind; on the contrary, it appears to be an easily remembered
exposition in which has been combined a high resistance to
action of reagents, a small tendency to devitrification, and
possibility of melting it and of working it into the finished
re. These are the criteria by which this glass has been ad-
iged to be a chemical compound, the excellence of its proper-
', as a glass.2

fhe second line of investigation, while scarcely more fruitful,
it least a more scientific method of attack, and results have
n obtained which, besides being of theoretical interest, are
much practical importance. Yet such work must deal with
stallizable substances and gives information, so far as chem-
composition is concerned, regarding only the crystallized
terial. Since little knowledge is at hand regarding chemical
nges which may take place during crystallization, it is im-
sible to draw conclusions as to the chemical condition of the
;d mixture or of the supercooled glass.
*his paper is a record, primarily, of the results obtained in an

See O. Schott, Dingier' s polytech. J., 216, 346 (1877); Weber, Ibid.,
349 (1879); G. Wagener, Ibid., 243, 66 (1882); E. Tscheuschner, Ibid.,
75 (1885); Schwartz, Chem. Zentr., 1886, p. 825; Mylius and Foerster,

, 22, 1 104 (1889).

See Asehc and Asche, "The Hexite-Pentite Theory," p. 253.

•s

Ji IURNAL < >F THE

examination of the systems \;i<> CaO SiOa and ,\;i.o Ba
SiOs by means of the refractive index.

Table i, however, contains data taken from the literature
the feldspar glasses and Table 2 similar data for the sysl
CaSiOa MgSiOs (glasses). These data are shown in grafl
form in Figs. 1 and 2. Ac-cording to our hypothesis, since 1
refractive indices of the feldspar glasses lie on a single straij
line, no new compounds are formed in this system. In the CaSil
MgSiOs mixtures, however, three straight lines are obtained
tersecting at two points, which is interpreted as indicating 1
formation of two new compounds, and from the position of 1
intersections, these compounds possess the compositions CaM
(SiC>2)4 and Ca,Mg(Si03)5. It is of interest that no indication
given of the formation of diopside, CaMg(Si03)2, although t
is the only double silicate which crystallizes from these m
tures.1 This system, therefore, apparently offers an excelli
example of chemical changes which may take place during 1
process of devitrification of a glass. It is worthy of note tl
one of these new compounds possesses the chemical compc
tion of the mineral tremolite, CaMg3(Si03)4. The specific gra
ties of these glasses give the same indications as the refract
indices as are illustrated in Figs. 1 and 2.

Experimental.

For the experiments described in this paper sodium carbona
barium carbonate, calcium carbonate and a very good qual
of glass sand were employed. The purity of each of the r

Table i.

Albite-Anorthite Glasses

Anorthite.
Per cent, by

volume.

Refractive index.
Na line.

Density

Obs.

Calc.

Obs.

Calc.

O.OO

1 . 4890

1 . 4890

2.382

2.382

31.85

I .5166

1 .5166

2.483

2.483

48.32

I.5308

1-5308

2-533

2.536

65.16

I -5452

1-5453

2-591

2.589

82.39

I . 5600

1 . 5602

2.648

2 644

IOO.OO

1 -5755

1-5754

2 .700

2 .7OO

dN/dp -

= 0.000864

dD/dp = 0

. 003 1 8

1 Allen and White, Am. J. Sci., 27, 1 (1909).

AMERICAN CERAMIC SOCIETY.

79

ueoo

'J J '00

'.5 tOO

I.SSOO

/S400

h

:*'#-

i^

X

15300

«JJJ

^

IS700

15100

P

(n*lj

i

1.5000

14900

14600

1.4700

1

1

?

■>

3

;

4

0

^

7

6

0

r

''

a

0

9

o Jnorfh/fe

70 60 SO 40 30

Fig. i. — Albite-Anorthite Glasses.

Fig. 2.— CaSiOj-MgSiOs

8o

JOURNAL OF THE

Table 2.

CaSiO, MgSiO,.

CaSiOi.
Per cent, by

volume.

(

Refracti

Na

a- index.

inc.

E

knsity

Dbs.

Calc.

Obs.

Calc.

O.OO

I .5801

I .5801

2.758

2.760

4.76

I

5823

I

5828

2-777

2.771

9-54

1

5853

I

5854

2.781

2.781

28.92

I

5960

1

5960

2.823

2 .822

38.76

1

6008

I

6008

2835

2.834

52.35

I

6073

I

6073

2.854

2 . 85 I

58.75

I

6105

1

6104

2.858

2.859

62.80

1

6l22

I

6123

2 .872

2.865

72.99

I

6174

I

6172

2.881

2.877

84.60

I

6224

I

6224

2.89I

2.89I

94-75

1

6261

I

6261

2.899

2.899

100.00

I

6280

I

6280

2.904

2.904

dN/dp.

dV/dp.

(0-26.

78%)

= O.OOO557

O

00225

(26

78

-81.45%)

= O.OOO482

O

OOI25

(81

15

-ic

»%)

= O

OOO362

O

OO086

materials was determined and on this basis the several mixtt
were carefully weighed out, mixed and fused in hard burned c
crucibles in the oxidizing atmosphere of a gas furnace. A
the glass was thoroughly fused it was poured out on cold i
plates, and after annealing, unless devitrification prevem
was broken up into small fragments. The refractive index
white light was measured with the aid of an Abbe refractome
using selected fragments of the glass which presented a smc
and flat surface of the original plate. With a little care in
lecting plane fragments of the glass, the values of the ref:
tive index for successive plates usually agreed to within one
two units in the third decimal place and the average of a nurr
of measurements was taken as the final value of the refrad

index.

The System Na20 BaO-Si02.

In order that all the compounds formed in a ternary syS
may be identified, it is necessary that several series of mixti
be made, their number depending upon the number of I
pounds formed. The system Na20-BaO-Si02 serves to il

AMERICAN CERAMIC SOCIETY.

81

the procedure. Mixtures of Na20 and SiOa were first made
These varied in composition from Na2O.Si02 to Na20.5Si02.
>nd these limits either fusion was difficult or only crystallized
ucts were obtained. This series indicates a compound of
NTa20.2Si02. The second series might well have consisted
ixtures of Na20.2Si02 and BaO.Si02, but in order that de-

\
\

Nc

,0-2'-

,02

b

^<

*

1.510

H

-.

IS 00

l

'•>

1.490

G
1

,v

s

\

1.480

\

\

1.470

\

V

\

b

Fig. 3.

1

J

v---

^

->-<

t.'a.U

BoO

^

£~o

-*"

ii

^

1

Fig. 4.— Na20.3.5Si02-BaO.Si02

JOURNAL OF THE

vitrification might be eliminated as far as possible, and that t
mixtures might approach more nearly to real glasses, Nsi
3.5S1O2 and NaaO.sSiOj were used as variables with BaO.9

In one series, Na20.3vSi02 and Ba0.2Si02 were the variabh
In each of these mixtures the formation of a compound w
indicated at molecular ratios for Na2() and Ba() of one to or
The final series was then made up in which (Na2O.BaO) at
vSiOo were the variables. This series indicates compounds
the formulas Na2O.Ba0.3Si()2 and Na2O.BaO.4SiO.;.

The data for these series are given in Tables 3 to 7, and a
shown in graphic form in Figs. 3 to 7. In Figs. 4, 5 and 6 the e
istence of a compound is indicated in which the molecular ral
of Na20 to BaO is one to one ; in Fig. 7 this ratio of bases is mai
tained constant while the ratio of bases to Si02 was variab
The results, as shown in Fig. 7, indicate compounds of the coi
position Na2O.Ba0.3Si02 and Na2O.Ba0.4Si02. No compoun
of this type, containing more than four molecules of Si02, appe
to exist in these glasses. The glasses possessing the compo
tion of these compounds de vitrify readily, the ease of devit
fication increasing as the quantity of Si02 decreases. The ser:
BaO-Si02 was not extensively investigated, chiefly because
the high melting point and the readiness with which these m
tures crystallize. Two glasses were, however, made, cor:
sponding in composition to BaO.Si02 and Ba0.2Si02, and,
though the refractive index measurements for these glasses m
subject to rather large errors, the indications are that Ba0.2Si

Table 3.

Na20-

-vSiO-..

Silica.

er cent, by
volume.

Refractive
index.

IOO.OO

( I . 464)

85

50

(1.4865)

80

55

I • 4950

78

00

I . 5000

76

75

I . 5000

70

70

I 5IIO

63

80

1-5137

54

20

I .52OO

Table

4-

Na20.3.5SiOr

-BaO.SiOi.

BaO.Si02

Per cent . by

Refractive

volume .

index

O.OO

I -4950 j

4-42

I ■ 5030

8.44

I 5I50

1538

1 5255

17.69

1 -5330

21.97

1 ■ 5440

29.80

I 5603

39.OO

I -5750

1 Extrapolated from Fig.. 6.

AMERICAN CERAMIC SOCIETY.

83

Table 5.
Na,0.3SiO,-Ba0.2Si02

Ba0.2Si02.

Per cent, by
volume

O.OO

6.82

18.OO

25 -70

28.20

33 20

42.40

52.20

63.10

74 ■ 70

86.20

1 00 . 00

Refractive
index.

5000

5054
5200
5310
5330

537<>
5488

5594
57i8
5847
5945
6090

Si <>■...

Per cent, by

volume.

Table 7.
Na,O.BaO-SiO>.

Table 6.

Na20.5SiOr

BaO.SiO,

BaO.SiOs.

Der cent, by

Refractive

volume.

index.

3 96

I .496

7

42

I

504

1 1

46

I

515

17

43

I

526

23

95

I

541

32

0

I

555

42

3

I

571

Refractive index.
I.6055

1 -5755
1 .5716
1.5684
15650
1 • 5603

.5488

•5407

1

a definite compound. The two ternary compounds, there-
i, correspond to double salts of the simple silicates Na20.2Si02.-
DSiOo and Na20.2Si02.Ba0.2Si02.

The System Na20 CaO SiO>.

n Tables 8 and 9 and in Figs. 8 and 9 are presented data for
soda lime glasses. The series Na20.3Si02-CaO.Si02, as il-
:rated in Fig." 8, indicates the formation of a compound in
ich the ratio of Na20 to CaO is two to three. Fig. 9 shows
series in which this ratio was maintained constant while the
lo of base to Si02 was varied. As will be seen from the figure,
y one compound is formed, 2Na20.3Ca0.7Si02. Ordinary

Jl >URNAL OF THE

y*

noo

is to

151,0

^^

k

c
&

I

la,0-

y

**

'

v

A

I.HO

V

-*■

Mo-,

0
0 IS

I

0, 9

0
0

?
t

0
0

3

7

0
0

4

i

c
0

5

0
0

40

40

70

50

SO

20

9OBoO2Si0

10 C

Fig. 5.

/(t(,0

lt?0

jj

/ffifl

,

|

^o"

No2C

'-find

^>

>

1500

f

J*

<

S

Ma20-SSi(k 90

Fir,. 6.

AMERICAN CERAMIC SOCIETY.

85

/.too

NoM

BaO

W ^

V

\

^

Na,0

BaO

tSiOt

I5b0

f

1.540

!

.*

x*

1.520

1

\
\
\

*

\

1.500

\
\

\
\

I4S0

\

.

\

\

O

50
NoiOBaO 50

PC

90 5,V2

10 O

Fig. 7.

>da lime glasses contain much more soda and silica than corre-
>ond to this formula. On the one extreme a typical glass rich
alkali of a low melting temperature possessing a composition
jproximately represented by the following formula: i2Na20.-
2a0.45Si02. Such a glass is, therefore, considered to be made
5 of a solution of

2Na20.3Ca0.7Si02 1 Mol.

Na20.2Si02 10 Mol.

Si02 18 Mol.

a the other extreme the high melting lime -rich glass of the
lormal" formula Na2O.Ca0.6Si02 is considered as being a som-
an of

2Na-.0.3Ca0.7Si02 1 Mol.

Na20.2Si02 1 Mol.

Si02 9 Mol.

86

JOURNAL OF THE

Table *

Xa,<USi

O.SiO

i

Per cent I

volume

iy Kelt ui i\ .
Illclrv

O.O

1 ,V><">

4 35

1 s< rf>4

9 .05

t . 5 11 5

14 (>c>

15172

20.98

1 524°

28 45

1 533°

37 40

1 5425

47.90
61 .50

1 5590
1 5796

Table <>■
2Nat0.3Ca( •

Sic,.

Per ci-iii by

\ olume

60. 50

64 85

68 25

73.60

Kt-fr;u tlve
index

1 5715
1 5645
1.5583

1 5425

It is of interest, although not considered to be of vital im-j
portanee, to compare these results with the melting-point data
of binary systems. According to Kultascheff,1 the melting-point
diagram of Na^SiOa-CaSiOa mixtures exhibits a maximum, in-
dicating a compound of the composition 2(Na2Si03).3(CaSi():i).
This was confirmed by the work of Wallace.2 Wallace has also!
investigated the system Na2SiO;j-BaSi03 and has obtained data,
which indicate that these mixtures form an unbroken series of
mixed crystals, and that no double silicate, therefore, separates
from these fusions.

The melting-point diagram of the system Na20-Si02 does not
appear to be on record. Investigations3 of "water glass" indi-
cate that Na20.2Si02, or some hydrate thereof, is the most acidic
silicate of sodium which is stable in aqueous solution. Reason-
ing from analytical data, Legorio,4 in a study of crystallization in|
eruptive magma, concluded that the compound Na20.2Si02
might be regarded as the common solvent in such silicate fusions,
and referred to it as the "glass base." Of greater significance, how-
ever, is the observation of Gelstharp5 that all mixtures of silica

1 Z. anorg. Chem., 35, 187 (1903)-'

2 Ibid., 63, 8 (1909).

3 Lielegg, Dingier' s polytech. J., 153, 44 (1859); Scheurer-Kestner, Rep.i
Chem. Applique, 5, 150 (1863); Scherer, J. prakt. Client., 41, 415 (186+ :
Kohlrausch, Z. phys. Chem., 12, 773 (1893); Mylius, Ber., VI Int. (
Appl. Sci., 1, 677 (1908).

4 Tscherm. Min. Mitteil., 8, 421 (1886).

5 Trans. Am. Ceram. Soc, 14, 647 (1912).

AMERICAN CERAMIC SOCIETY.

87

/MO

^

>

/ISO

. /

'

1360

l

k

£'

V

1

2Nc

<,0 1

CaQ

P*

/.540

fc

<i?'

1.570

0

s

y

1.500^

s

>

0

10

?0

30

40

SO

60

70 CaO.S'O.

03

5/0/ 90

SO

70

60

50

40

30

Fig. 8.

1.580

1

1.5 tO

0 75,

4

N

2Na,

1.540

\

\

\

1.520

\

,

\
\

1.500

1

\
\

N

\

\

1.480

\
\
\

\
\

1 4b0

SO 60 70 80 90 5,0.

ZNo,OlCoO &0 30 20 /0 0

Fig. 9.

88 JOURNAL OF THE

and sodium carbonate, which contained less silica than com

s|)oii(ls to the above formula, resulted in clear fusions when tha
reaction was complete, as evidenced by the cessation of the evolu-
tion of carbon dioxide. When, however, the quantity of silica
was greater than that necessary to form Na2O.2vSi02, undissolved
silica was observed after the chemical reaction had gone to com-
pletion.

COMMUNICATED DISCUSSIONS.

C. II. Kerr: All who have undertaken the study of any phase
of equilibrium in silicate melts appreciate the great difficulties
involved and any method which gives promise of opening up
new lines of attack is exceedingly welcome. It seems that the
present method offers great possibility of development as a new
means of studying the problems of glass manufacture and it
should prove equally valuable in other fields where the reacr
tions are carried well on toward equilibrium.

In the present investigation, it is obvious that the whole field
of non-de vitrifying glasses, which can be made from the ingre-
dients herein describe d, has not been covered, but it is not shown
in the paper just what relation the compounds which have been
studied bear to the entire field. It seems that a much more
thorough investigation, at much closer composition intervals
but embracing necessarily a greatly restricted number of composi-
tions, would have yielded results of even greater value. There
are not sufficient points on any one curve to show with any
great assurance of correctness the nature of the curves and their
significance.

In the context it is stated that the index for white light was
determined. It is well known that the Abb6 refractometer
gives readings for the D line (or sodium line) of the spectrum.

As to the method of making index determinations, the speci-
mens described are not considered to be satisfactory for the work
in hand. The surfaces produced by pouring a glass melt onto
an iron plate fall far short of being sufficiently good for readings
of index in investigative work. Variations in index readings,
due to surface irregularities, wTill, in many cases, exceed the varia-

AMERICAN CERAMIC SOCIETY. 89

Dns due to the comparatively small changes in composition,
ata are not submitted in detail showing just what variations
i index were found in the individual readings on the same melt.
: is a well-known fact that, even with a highly polished piano
;st piece, a very slight tipping up of one edge, due to a minute
ist particle, will cause errors as great as one unit in the third
primal place, and surface irregularities of an ordinary fire pol-
hed specimen will show variations in index readings up to one

even two units in the third decimal place. The surface pro-
iced by pouring the melt onto an iron plate will probably cause
ren greater variations and therefore, with any reasonable num-
:r of readings, the variation due to surface is apt to be quite
rge. This must be borne in mind throughout the present
:>rk.

It is regretted that full tabulations of all readings are not re-
>rded or, at least a full tabulation in one case, to show the actual
athematical probabilities. When it is realized that variations

one or two units in the third decimal place may occur in the
dividual readings, and probably errors as large as that due
ily to the irregular surface of the specimen, and still at the
me time variations of a few units in the fourth decimal place
e taken to indicate departure from a straight line, hyperbolic

other simple curve form, it is evident that a very careful con-
leration must be given to each factor influencing each deduc-
3n before it can be considered as established that the existence

definite compounds is shown.

Mr. E. D. Tillyer, of the American Optical Company, has
ven some consideration to the mathematical side of the problem,
aking the data in Table 2 and using the Cornu interpolation
rmula, he says that the probable error in the assumption of
lis formula (that an hyperbolic curve will fit various points
1 the chart) is substantially the same as the probable error from
e assumption that three intersecting lines accurately repre-
nt the facts as in Fig. 2, given by the author. In other words,
1 hyperbolic curve, drawn through these points, is of the same
der of probability in its correctness as are the three intersecting
raight lines and, obviously, the hyperbolic curve would indi-

9o I1 '1 RNAL OF THE

cate nothing, so far as we arc- now aware, about the existence
of anv definite compounds. Furthermore, the hyperbolic curve
is the simplest physical expression we can conceive to indicate
progressive changes in index of refraction or in other wave-length
manifestations. 1 am greatly indebted to Air. Tillyer for his
consideration of these points.

These facts are not presented to show that the author's conclu-
sions are in error, which is neither claimed nor denied, but merely
to direct careful study regarding the mathematical significance
of the data presented.

( >ne further point to which attention should be drawn is the
desirability of making greater use of the specific gravity deter-
minations. The Lorenz-Lorentz formula has demonstrated that
specific gravity changes must necessarily accompany changes
in the indices of refraction. It is well known that the accuracy
obtainable in specific gravity determination is several times
the accuracy of determinations of the indices of refraction and,
therefore, a study of the specific gravity variations in such con-
nections as the present appears to offer very attractive possi- :
bilities.

A. E. Williams, C. C. Rand: It has been our experience in
making experimental glass melts, even in quantities as small as
200 grams, that the glass cannot be made homogeneous enough
to obtain an index of refraction which is more accurate than one
point in the second place, by simply melting the mixture thor-
oughly. The glass must be stirred for a considerable time be-
fore an index correct to the third place is possible. Further-
more, the change of index due to volatilization is considerable
and may cause a difference of one point in the second place be-
tween an experimental melt and a iooo-pound melt of identically
the same batch composition.

If the compositions of the glasses for which the indices and
densities are given are the batch compositions rather than the
analyses after melting, we believe that the variations in the
curves are within the limits of error possible from unhomogeneity
in the glasses and changes of composition due to' volatilization.

AMERICAN CERAMIC SOCIETY. 91

E- W. TiLLOTSON: The questions raised in these discussions are
hose \\hich naturally present themselves to those who have
/orked in this field. It is to be recognized that some error is
iresent by reason of lack of homogeneity, and possibly as a
esult of volatilization. It was the purpose that errors from
hese sources were to be minimized by carrying out each experi-
nent, as far as possible, under the same conditions. While the
alues for the indices given iray not be absolute values and may
iffer from those secured in larger melts, it is believed that they
re comparable among themselves within the limits mentioned.

The method of making the index determinations was, perhaps,
ot presented clearly. While the glasses were poured on iron
lates, it was the upper "fire polished" surface which was used
:>r making the measurements. By breaking the plate several
mall pieces could be obtained, with surfaces suitable for the
fteen or twenty readings taken.

With regard to the theory of the "straight line" relationship,
ie following illustrates the mathematical considerations:

Let «i, nu h: U2, n-2, h, etc., represent the velocity of light,
tractive index and time required for light to traverse the dis-
mce /1, h, etc., of the several components, and L, U and T be
le corresponding quantities for vacuum. The following rela-
onships then follow:

- = n-2 + etc.

+ etc.

u

= nx

u

14-2

= Wo

Hi

h

_ I,
Mo

T = L
U

«2

h

" T

.. = - + etc.

If now the mixture be considered to be made up of an infinite
umber of layers through all of which the light must pass con-
:cutively, and since

L = h + k + etc.,

le time required for light to traverse the mixture will be

JOURNAL OF THE

h + k + etc.
The refractive index of the mixture will then be

-ft)g) +(»©+-*

Substituting for , , etc., their equivalents nu n^, etc., and for

Mi W2

— , — , etc., the percentages by volume ph pi, etc., the equa-
te J-/

tion becomes

N = piH] + pin-y + etc.

This indicates, therefore, that in mixtures, in which no new
compounds are formed, the refractive indices would be expected
to follow the "straight line" law, when the composition of the
mixtures is expressed in percentages by volume. When new com-
pounds are formed quantitatively in the mixture, it is to be ex-
pected that the linear relationship wil still hold, but the indices
may lie on several straight lines which intersect at compositions
corresponding to those of the new compounds.

It may be added that nearly all of the data here presented was
secured before the linear relationship between index and composi-
tion by volume became apparent. However, before it was
adopted, confirmatory evidence was sought in published data
on solutions and on mixtures of liquids. Nearly one hundred
systems were recomputed on the volume percentage basis, and
many systems were found in which the linearity is so striking
as to make it difficult to believe that it is accidental. It is also
difficult to explain, by the laws of chance, that the data should
lie so nearly on straight lines which intersect at compositions
corresponding to simple molecular ratios. If the true "curve"
is hyperbolic or if the errors of experiment are so great as to
mask the true curve, it is remarkable that the indicated linearity
should present itself with such regularity. Incidentally, the writer
is not aware that he has taken "variations of a few units in the
fourth decimal place to indicate departure from a straight line."

AMERICAN CERAMIC SOCIETY. 93

With reference to the use of interpolation formulae, it is to
be recognized that any "curve" may be approximated as closely
as may be desired by the judicious selection of a formula of a suffi-
cient number of terms. The fact that such an empirical expression
may approximate a given series of points does not militate against
a second expression which is founded upon a theoretical con-
sideration of the problem and which gives an equally good agree-
ment with the series of points in question. The hyperbolic
function may be the most simple expression relating changes
of refractive ndex with varying wave lengths or with varying
temperatures, but it is not apparent that it should relate change
in the index with changes in composition.

The statement that "specific gravity changes must necessarily
accompany changes in the index of refraction" is true only under
certain conditions. The formula of Lorenz-Lorentz relates
the index with the density of a substance of given composition,
under different conditions of temperature. In the systems
under consideration, the temperatures remain substantially
constant while the composition varies. It may be true that a
glass of the highest density will have the larger index, but this
is not universally true of mixtures in general. Mixtures of
water and pyridine, for example, illustrate the reverse condition.

25

D25" Nd. 25.2°.>

Water 1 .000 1 . 33266

Pyridine o . 976s 1 . 50677

In this system the mixtures of the lower density have the higher
index.

It is the writer's belief that the data here presented indicates
a new line of attack for investigations on the constitution of
glasses and other solutions, and it is to be hoped that others
may subject the principles indicated to a minute experimental in-
vestigation.

Mellon Institute,
Pittsburgh, Pa.

1 Zawidski, Z. phys. Chem., 35, 129 (1901).

2 Van Nostrand's Chemical Annual.

FIRE CLAYS IN NORTHERN IDAHO.

Bv S. K. Sopbk, Corvallis, Oregon.

Location and Accessibility. During the past three years
there have been several important discoveries of high grade fire
clay in Latah County, Idaho. Latah County is in the northwest
part of thf State, adjoining Whitman County, Washington.
The clay deposits occur near Moscow, the County scat, and
site of the University of Idaho. The deposits lie in close proximity
to the Lewiston branch of the Northern Pacific Railway, and at
the property of the Moscow Fire Brick and Clay Products Co.,
near Moscow, a railroad spur is being built to the clay pits to
permit of the loading of the clay direct from the pit to the cars.

Geology of the Clay Deposits. — The geological occurrence
of the clay deposits is interesting and some rather unusual fea-
tures are exhibited. The fire clay is of the residual type, and
has been derived from the weathering and decomposition of
granite and associated pegmatite dikes. In addition to these
fire clays there are other residual clays derived from the decom-
position of basalt, which occurs over a large area in this region
as a capping on top of the granite. The clays derived from the
decomposition of the basalt are red, yellow or brown in color.
Because of the higher percentage of iron and other fluxing im-
purities, they have little value for refractory purposes, but are
excellent for the manufacture of brick, tile and other structural
clay products. The fire clays are usually white or cream col-
ored, and are white burning.

The general geologv of this section of Idaho is comparativelv
simple. The rocks in the vicinity of Moscow consist of granite,
quartzite, and basalt. The quartzite, which is the oldest rock,
probably Algonkian in age, has been altered to a quartz schist
in places. The granite occurs as a large intrusion into the quartz-
ite and quartz schists and is probably of Cretaceous age. It
forms the edge of the enormous granite batholith which, be-
cause of erosion, is now exposed at the surface over many square
miles in the mountainous region to the east and southeast. The

AMERICAN CERAMIC SOCIETY. 9.5

quartzite and associated pre-Cambrian sediments have been mostly
eroded from the region under discussion, but large areas are ex-
posed to the northeast.

The basalt occurs as a flow, or series of flows, which originally
covered the entire region of the Snake River plains in Idaho,
Eastern Washington, and a large part of Eastern Oregon.
This formation is known as the Columbia basalt and is
probably of Miocene age. The lava flows originally reached
well up on the foothills and lower western flanks of the moun-
tains to the east of Moscow. Subsequent erosion has removed
the basalt capping from many of the upper slopes of the foot-
hills bordering the mountains and it is along the contact of
this basalt cap-rock and the denuded granite that the fire
clays and associated basaltic residual clays occur. Although
there are large areas of granite, just east of this contact, from
which the basalt has been completely removed, no important de-
posits of fire clay have been discovered as yet on these terrains.

The occurrence of the clay along the contact zone is probably
due to the fact that on the granite side of the contact the hills
rise to considerably higher elevations with rather steep slopes
and the clay and decomposed granite is rapidly removed by
erosion. Along the contact zone the slopes are more gentle
and thus clay deposits of considerable thickness have accumula-
ted. Another probable reason for the close association of the
clay deposits and the zone of contact between granite and basalt
suggested itself to the writer in examining one of the recently
exposed clay banks near the town of Troy. At this locality the
clay is overlain by a bed of volcanic tuff (consolidated volcanic
ash), which, in turn, is partly overlain by basalt. At this place
the tuff is exposed in immediate contact with the underlying
residual fire clay which grades downward into partially decom-
posed granite. For a thickness of several feet beneath the tuff
the clay is stained dark brown or yellow and contains numer-
ous altered fragments of wood, showing plainly the existence
here of an old weathered surface antedating the volcanic ac-
tivity which resulted in the deposition of the ash. This evi-
dence points to the conclusion that in some cases the residual
clay was formed before the lava flows which now cover so much

96 JOURNAL OF THE

of the region. Recent erosion of the lava capping lias resulted
in the exposure- of tin- clay and decomposed granite and this
in turn has lircii rapidly removed because of its relatively weaker
resistance to erosion processes. The more rapid erosion of the
areas of decomposed granite would tend to remove the clay ex
cept along the zone adjoining the contact, where the protecting
capping of basalt, or tuff, has not been removed sufficiently
long to result in the erosion of the clay. It is probable that
much of the fire clay is Miocene or pre Miocene in age, but as
considerable decomposition of the granite has taken place since
the erosion of the basalt, this has probably resulted in the forma-
tion of post- Miocene residual clay deposits.

The red, yellow, and brown clays which have been derived
from the decomposition of the basalt are clearly of recent age.
As would be expected, these generally occur at somewhat higher
elevations than the white fire clays and in some cases are found
directly overlying the fire clay. Recent erosion has resulted
in the accumulation of secondary deposits of mixed sedimen-
tary clays derived from the removal and mixing of the residual
fire and basaltic clays.

The granite from which the fire clay in this locality has been
derived is rich in feldspar, and generally contains only small
amounts of hornblende, biotite or other .ferro-magnesian min-
erals. In the vicinity of Moscow, the granite shows almost no
mica and hornblende. Considerable white mica occurs in the
granite at some places and there is also a segregation of the feld-
spars along narrow zones cutting through the granite in various
directions. This has resulted in the occurrence of numerous
narrow streaks of almost pure kaolin cutting the clay deposits.
Narrow pegmatite dikes are also of rather frequent occurrence
in the granite of this locality and this pegmatite has contributed
additional amounts of kaolin, to the fire clay. Because of the
numerous segregations of the feldspar, white mica, or quartz
constituents of the granite, the material resulting from the de-
composition of the granite is often "spotty" in its occurrence.
In places this material will consist almost entirely of sub-angular
quartz sand. In other places, large amounts of partially de-

AMERICAN CERAMIC SOCIETY. 97

composed mica were noted, especially at the O. K. Olsen mine
near Troy. In some places, where the proportion of feldspar
in the original granite was high, deposits of good white fire clay
have resulted from the weathering of the rock.

Extent of the Clay Deposits. — The size and shape of the clay
deposits are variable. They range in thickness from a few feet
to 40 or 50 feet, and the area of the deposits may vary from a
fraction of an acre to 40 or 50 acres. This variation in size and
shape is to be expected from the method of origin of the deposits,
and from the character of the parent rock. In prospecting or
developing these deposits, it is advisable to put down a large
number of bore holes or test pits, arranged on some systematic
plan, in order to obtain accurate data from which the extent of
the deposit may be outlined.

Quality of the Clay. — The fire clay is of excellent quality
and is well suited for the manufacture of fire brick. The best
and purest deposits are those being worked at Moscow. Heat-
ing tests have shown that the clays have softening points ranging
from that of Cone 30 to Cone 36. As the material comes from
the pits, it develops only a moderate degree of plasticity and is
of coarse texture, containing numerous angular grains of quartz
embedded in white kaolin. There is sometimes present small
amounts of white mica but in the Moscow deposits the quantity
of mica in the clay is negligible. The kaolin is often rather
dry and powdery, which makes the clay crumble easily, but
when water is added and the material thoroughly mixed, the clay
develops a surprising degree of plasticity. Up to the present
time, no attempts have been made to utilize these clays for any
purpose other than in the manufacture of fire brick and other
refractory shapes. However, if the clay were washed, kaolin
and silica of very good quality should result. The former might
be utilized in the manufacture of pottery and other high-grade
wares, and the white quartz, which is of unusual purity, might
be of value in the manufacture of silica brick.

Development and Market. — The discovery of the fire clays
of this- locality is of recent date, but already two plants are in

AMERICAN CERAMIC SOCIETY.

operation manufacturing refractories. The plant of the Mos-
cow Fire Brick and Clay Product Company at Moscow, Idaho,
was placed in operation in i<)i-j and from the start this company
has been operating continuously. Another plant is in operation
at Troy, Idaho, about 10 miles east of Moscow. There are also
brick plants at Mica and Freeman, in Spokane County, Wash-
ington, northwest of Latah County, Idaho. These Washington
plants are operating on clays of similar origin to those here de-
scribed. The market for fire brick and other refractories in the
Northwest is very encouraging. The large smelters at Kellogg,
Idaho; Northport, Washington; Tacoma, Washington, and in
British Columbia, furnish a regular market for large quantities
of refractories in addition to the amounts regularly used in the
various industrial works at Spokane, Seattle, Tacoma, Portland,
Vancouver, and other points in the Northwest.

It is not unlikely that a steel plant will be erected on the north
Pacific coast in the near future, affording an additional market
for fire clay and other refractory products. The Moscow Fire Brick
and Clay Product Company is also utilizing deposits of the residual
basaltic clay, which occurs on their property, in the manufac-
ture of brick, tile and special shapes. Several varieties and colors
of this tvpe of clay occur in the Moscow deposit, and by their
mixture with variable quantities of the fire clay, a wide range
of red, buff, brown, and spotted brick and other products can
be produced.

Oregon State School of Mines,
Corvallis, Oregon.

GROUND COAT ENAMELS FOR CAST IRON.1

By Homer F. Stale y, Ames, Iowa.

The function of a ground coat enamel is that it should serve
as a bond between the cover enamel and the iron and also pro-
tect the iron from oxidation while being heated to the tempera-
ture at which the cover enamel fuses. Statements are some-
times found in the literature that a ground coat is not necessary
when powdered enamels are used for cast iron.2 While it is some
times possible to enamel a trial piece without the use of a ground
coat, the percentage of good pieces obtained in this way is very
small and the process is not a commercially feasible one. At
the present time ground coats are universally used in the enameling
of cast iron in this country.

It was formerly the practice to use very refractory ground
masses which were merely sintered onto the iron but were not
fused to a glass. The object was to produce a porcelain-like coat-
ing on the metal, toward which the cover enamel would act as a
glaze. This type of ground coat is still used for wet-coat enamels
on cast iron but for the powdered enamels it has been superseded
to a large extent by thin glossy ground coats.

The basis of these sintered ground coats was a frit prepared
from flint and borax, or from flint, feldspar and borax, with small
additions of lead or sodium oxides. To this frit, clay and flint,
or clay and feldspar, were added in sufficient quantities to make
the refractoriness of the mass such that it would sinter, but not
fuse in the enameling oven. Magnesium oxide or sulphate was
used in small quantities to assist in floating the enamel. The
ground coat was fired until it could not be rubbed off with the
fingers and until the individual grains appeared rounded when

1 By permission of the Director, Bureau of Standards.

2 Holderoft, H., Jour. Soc. Chem. Ind., 29, 123.

[oo JOURNAL OF THE

examined with a good hand glass.1 Formula No. i is a ground
coat of this type which lias been used in tins country in the prepara
tion of wet coat enamels. Further examples can be found in
Randau, pages i_'4 and [25.

In the use of glassy ground coats, enamel makers have dis
carded the idea of there being an analogy between the enameling
of cast iron and the glazing of porcelain and have attempted to
produce glasses that would afford the maximum adhesion be-
tween the enamel and the metal. A satisfactory ground coat of
this class should melt at a dull red heat, in order to protect the
iron from oxidation; should be able to dissolve any oxides or
foreign matter on the surface of the metal; should be sufficientlv
fluid to flow in part, into the minute pores of the metal, so as to
produce a good bond, and should not blister or volatilize (burn
off) until temperatures above those commonly employed in
enameling furnaces are reached.2

The mixtures commonly employed in preparing the frits for
this type of enamel are flint or sand, feldspar, borax, red lead
and sodium nitrate. Flint and feldspar are the refractory in-
gredients of the enamel and supply the silica essential to the
production of a permanent glass. Flint or sand is often used
alone as the refractory component, especially in the older for-
mulae, but many recipes call for the use of some feldspar in ad-
dition to the flint. Borax is used in the largest proportions as
a flux on account of the ability of boric oxide to dissolve iron and
other oxides. A ground coat glass high in boric oxide readily
dissolves any small amounts of iron oxide present on the surface
of the iron. Lead oxide is used because its compounds melt at
low temperatures and produce fluid glasses having good mechan-
ical strength when cold. The small amount of sodium nitrate
is introduced in order to secure the oxidizing effect of the nitrate
radical which prevents the reduction of the lead and the absorp-
tion of injurious sulphur gases by the glass during the fritting
process. Small amounts of other fluxes such as fluorspar, cryolite,

1 Paul Randau, "Enamels and Enameling," 2nd English edition, pp.
20 and 148, Scott Greenwood & Co.

2 Trans. Am. Ceram. Soc, 13, 531 (1911).

AMERICAN CERAMIC SOCIETY. 101

arium oxide, soda ash, etc., are sometimes used, but these per-
>rm no distinctive service and most of the enamel frits are com-
ounded from the above list of materials. Small amounts of
lagnesium carbonate and magnesium sulphate are sometimes
sed in enamel frits on the assumption that they aid in the ad-
esion of the ground coat to the iron.1 Cobalt oxide in small
nounts is employed in many enamel frits for the same purpose.

As a raw material to be added at the mill, clay is invariably
sed on account of its ability to cause the enamel composition
I a whole to remain in suspension. It also serves as a refrac-
>ry ingredient and, in some cases, is the only rawr refractory
aterial introduced. Flint and feldspar are occasionally used
; raw refractories. Magnesium carbonate, magnesium sulphate,
>rax, lime water, ammonia, etc., are used in small amounts to
isist in floating the enamel. Sometimes a little raw cobalt oxide

added in cases where it has not been convenient to include in
le frit.

The function and value of cobalt oxide in ground coats for cast
on is a debatable question. After a careful investigation of the
ibject, Coe came to the conclusion that "the use of cobalt oxide

a ground coat for cast iron enamels is of doubtful value."2 ■ On
te other hand, manufacturers of sheet steel enamels in general
aim that they cannot produce a satisfactory ground coat with-
it the use of cobalt oxide or some other metallic oxide as a sub-
itute for the cobalt. The remarkable effects claimed for the
sry small amounts of cobalt oxide added are not accounted for,
id various unproven theories are advanced in an attempt to
plain the phenomena.3

Since many excellent ground coats contain no cobalt, it is evi-
:nt that cobalt oxide is not essential to the production of a
tisfactory ground coat for cast iron. On the other hand, various
ctors enter into the production of satisfactory enameled iron
ire and it is very difficult to determine the effect of the pres-
ce of cobalt oxide in the ground coat. As a result many

1 Paul Ratulau, "Enamels and Enameling," p. 126.
- Trans. Am. Ceram. Soc, 13, p. 545 (191 1).
3 R. D. I andrum, Ibid., 14, pp. 756-763 (1914).

102 JOURNAL OF THE

enamelers, including the writer, in some instances use cobalt
oxide in the preparation of ground coats because they feel thai
it is not detrimental and may possibly be beneficial. The
of the cobalt oxide, compared to the value of the ware produced,
is so small that its elimination from the formulas is not ad-
visable.

As demonstrated by Krnest Mayer for glazes,1 the settling of
enamels is due to an excessive alkalinity of the enamel solution.
Mayer divides the materials used for the flotation of slips into those
whose action is physical and those whose action is chemical. In
the first group are clay, syrup, gum arabic, dextrine, milk and blood.
The two latter soon curdle and make quite thick suspensions.
In the chemical group are various acid substances, including boric
acid, vinegar, hydrochloric and sulphuric acid, which reduce
the alkalinity by neutralizing a part of the alkali. A more ef-
fective group of chemicals are those which neutralize a part of
the alkalinity and at the same time produce a flocculent precipi-
tate. The most commonly employed are magnesium sulphate
and magnesium chloride :

MgS04 + 2NaOH = Mg(OH)2 + Na2S04
MgCl2 + 2NaOH = Mg(OH)2 + 2NaCl

It is a noteworthy fact that not one of these chemical vehicles
has any appreciable coagulating effect unless clay is present.
It would appear, therefore, that the effect of these reagents is to
coagulate the clay and thus render it capable of floating the
other enamel ingredients. By the addition of acid substances,
the enamel suspension would be rendered acid or faintly alkaline,
a condition which is favorable to coagulation. By the addition
of magnesium salts, the alkalinity would be reduced, a flocculent
precipitate. Mg(OH)2, would result, and a coagulating salt, sodium
sulphate or chloride, would remain in solution. Ammonium
carbonate is a very effective coagulant while borax serves as a
coagulant or deflocculent according to the conditions. The other
chemicals used for the floating of enamels are capable of flocculat-

1 Trans. Am. Ceram. Soc, II, 369 (1909).

AMERICAN CERAMIC SOCIETY. 103

jing the clay under the proper conditions and we have every reason
for assuming that this is their mode of action.1

As pointed out by Mayer, it is theoretically a mistake to intro-
duce a sulphate, such as magnesium sulphate, into a ground coat
on account of the danger of the development of sulphur blisters.
However, magnesium sulphate is the most commonly employed
vehicle with the exception of clay, and apparently the small
amounts added are not harmful. Greenwald recommends the
addition of magnesium oxide (calcined magnesium carbonate)
and ammonium carbonate as being chemically harmless.2

Magnesium oxide and carbonate, calcium oxide, calcium hy-
droxide (milk of lime) and calcium carbonate, may cause serious
difficulties if deposited upon the ware in large granules. These
granules will not be melted into the ground coat or enamel but
burn to particles of quicklime (CaO or MgO). In the course
of several months some of these particles will hydrate, the mois-
ture presumably working through the porous iron with which
they are in contact, and expand. This causes the enamel above
them to break off in little cone-shaped spalls. This liability is
not an imaginary one, for the writer has inspected ware to the
value of twenty thousand dollars, in one stockroom, ruined in
this manner. For this reason it is preferable to eliminate the in-
soluble compounds of magnesium or calcium as vehicles for float-
ing the enamels.

The so-called "vehicles" may be added when the ground coat
is charged into the mill, but the more common practice is to add
the clay just previous to the grinding and to add the others to
measured quantities of the ground coat just before its application.
When vehicles other than clay are used, the consistency of a ground
coat varies greatly with age.

The most simple and oldest type of glassy ground coat consists
of a refractory frit, high in sand or flint and generally containing
cobalt, to which only enough clay to float it is added at the mill.
Ground coats Nos. 2 and 3 are typical formulas, showing the
range of lead oxide and borax commonly employed in this type

1 H. E. Ashley, "Technical Control of the Colloidal Matter in Clays,"
Bur. Standards, Tech. Paper, 23, 74-102.

2 Sprechsaal, 43, 594.

i<»4 JOURNAL OF THE

of j^roimd coat. The variations is cobalt oxide are also within
the normal limits, ('.round coats containing a high percentage
of borax are more popular than those high in lead oxide, and
formulas similar to No. 2 are in more common use than those
similar to No. 3.

With the general adoption, in late years, of feldspar cover
enamels, it is quite natural that feldspar should have been intro-
duced into ground coat frits. The feldspar has not replace!
flint and sand entirely as in many cover enamels — a formula
for a ground coat containing neither flint nor sand being very-
unusual.

With the introduction of frit kilns it was not advisable to in-
clude the cobalt in the ground coat frit, as the kiln could not be
used for both the white and colored enamels. Consequently, in
some modern formulas we find cobalt oxide as an addition to
be made at the mill. No. 4 is a typical ground coat formula
containing feldspar in the frit, the cobalt being added at the
mill.

It was quite natural that, from time to time, enamel mixers
should attempt the blending of two ground coats. In some cases
the results were so satisfactory that the use of a formula calling
for the blending of twro ground coats became an established fac-
tory practice. According to the results obtained by J. H. Coe,1
in some cases the blending of two ground coats having about the
same heat range will produce coats having longer heat ranges
than is obtained by the use of either coat alone. No. 5 is a
typical ground coat formula of this kind. Incidentally it calls
for the addition of raw magnesium carbonate.

In the four glassy ground coat formulas already given, the
frits themselves are quite refractory and comprise the bulk of
the coat. The raw material consists essentially of only sufficient
clay to float the enamel. In late years, since the rather general
use of frit kilns for the smelting of ground coat frits, the use of
formulas containing more fusible frits and larger amounts of re-
fractory raw additions has been practiced in a number of plants.
Formula No. 6 is typical of a number of ground coats in use

1 Trans. Am. Ceram. Soc, 13, 531-549 (1911).

AMERICAN CERAMIC SOCIETY. 105

rhich contain large amounts of clay as the raw refractory addi-
ion., One advantage of this type of formula is that the use of
hemical "vehicles" for floating the coats can be avoided.

The use of clay alone as the raw refractory, in case the amount
o be added is large, is not as common as the employment of a
mall or moderate amount of clay as a floating agent and the ad-
[ition of either flint or feldspar, or both, as refractory ingre-
[ients. No. 7 is a formula of a ground coat calling for clay, flint
nd feldspar as the raw additions. In this formula two frits are
•lended, the effect being the same as the blending of two ground
oats. This ground coat is more fusible than the preceding ones,
leing employed in the enameling of light castings.

It is somewhat difficult to estimate the comparative fusibilities
>f a series of enamel formulas, whether these be presented in the
•roDortions of the raw batch, percentage composition, or empirical
hemical formulas. When only one refractory ingredient is used,
ome idea of the fusibility may be obtained by observing the
»ercentage of the refractory addition in the melted enamel. For
his reason, in submitting the composition of cover coat enamels,
le have employed formulas calling for potash feldspar as the
>nly refractory. It is impossible to follow this practice in ground
oat enamels, for the clay, flint and feldspar must be taken into
onsideration.

By practical trial the writer has determined that, in cast iron
namels, the potash, feldspar, flint and clay may be substituted
or one another without changing the fusibility of the enamel,
he ratio being 100 feldspar : 66 Vs flint : 40 clay.1 By assuming
hat flint is substituted in these ratios for all of the clay and feld-
par contents in each of the ground coat formulas given here
nd by reducing the batch to 1000 pounds (melted), we obtain

number indicating the pounds of flint that would have been
sed for 1000 pounds of the ground coat melted, if the only re-
ractory employed had been flint. We have called this number
he "flint equivalent" of the formula. The flint equivalent of

ground coat gives a fair indication of its refractoriness, for
ariation in the amount of the flint added has such a decided effect

1 Trans. Am. Ceratn. Soc, 13, 505 and 534 (191 1).

[o6

loi'KNAI, OF THE

thai the influence of variations commonly found in the relative
amounts of the fluxing oxides may be neglected.

('.round coat No. i is quite refractory and has a high "flint
equivalent." ('.round coat No. 7 is quite fusible and has a corre
spondingly low "flint equivalent." ('.round coats Nos. 2 to 6
are supposed to be well adapted to general lines of enameled iron
wares. It is remarkable how nearly uniform are the "flint equiva-
lents" of these five recipes obtained from various sources.

Ground Coat No. 1.

Frit

Flint

Borax

Sodium nitrate
Red lead

Hatch for 1000 pounds
Raw

Additions.

Flint. . .
Clay.. . .

Melted.

3.S'> o
[85 .0

35 "
41 .0

611 .0 500.0

345.0 3450
180.0 155-0

1 136.0 1000. o

Sit)..
AM I
Na,( )
B208. .
PI O..

Percentage
composition.

84.99

"Flint equivalent" =

Empirical Chemical Formula
0.807 AI.O3

0.785 Na20
0.215 PbO

14.743 S1O2
1 . 103 B203

AMERICAN CERAMIC SOCIETY.

107

Ground Coat No. 2.

Batch for 1000 pounds.

frit.

lint

lorax

odium nitrate . .

Led lead

Cobalt oxide. . . .

idition.

:iav. . .

Flint equivalent"

Raw. Melted.

675.O

390.0

35-0

52.5
30

"55-5 948.0

Si02- •
A1203-
Na20 .
B203.
PbO..
CoO.

60.0 52.0

1215.5 1000. o

= 736

Empirical Chemical Formula.
0.810 Na20 ] f 7.705 Si02

0.165 PbO } 0.158AI2O3 { 1.342 B2O3
0.025 CoO J

Percentage
composition.

70.28

2-39

7.60

14.27

5-15
0.30

99 99

72.67

Ground C6at No. 3.

Batch for 1000 pounds.

frit. Raw. Melted.

lint 680.0

orax 1 70 . o

Ddium nitrate. . . 50.0

ed lead 170.0

obalt oxide o 85

1070.85

ddition.

lay 51.0

956.O

44.O

1121.85 IOOO. o
Flint equivalent" = 735

Percentage
composition.

Si02 70.36 \

AI2O3 2.03 J72'39

Na20 4.57

B2O3 6.26

PbO 16.66

CoO 0.085

99-97

Empirical Chemical Formula.
0.448 Na20 1 [7.102 Si02

0.485 PbO [0.121AI2O3 J0.539B2O3
0.067 CoO

IOS

JOURNAL OF THE

Ground Coat No. 4.

Batch for 1000 poundi

SiOj..

A1203 .

KiO .

Na20.

B203.

PbO.

CoO.

Prit.

Raw.

Mi

Sand

275 0

Potash feldspar...

435 0

Borax

375 -o

Red lead

41 .0

Additions.

Clay

Cobalt oxide.

I 126.0 948.0

57 -o 49.o
3-o 3.0

1186.0 1000. o
"Flint equivalent" = 733

Percen

compos

1 1 loll

58
IO

30 \

.27 J

68.

7

35

6

.07

'3

72

4

< K 1

0

30

10a

.OI

Empirical Chemical Formula.
0.392 K20 f 4.884 Si02

0.493 Na20 \ 0.507 AI2O3 \
0.095 PbO J [ 0.985 B203

0.020 CoO

Ground Coat No. 5.

Blue ground coat.
Batch for 1000 pounds.

White ground coat.
Batch for 1000 pounds.

Frit.

Sand

Borax

Sodium nitrate.

Red lead

Cobalt

Raw. Melted.

685.O

335 -o

30.0

85.0

2.7

Frit. Raw.

Sand 395 .0

Potash feldspar 315 .0

Borax 335 .0

Sodium nitrate 30 . o

Red lead 60.0

Melted.

II37.7 958.0 II35-0 958.O

Additions. Additions.

Clay 48.0 41.0 Clay 48.0 41.0

Magnesium carbonate 3.0 1.5 Magnesium carbonate 3.0 1.5

.7 1000.5

1186.0 1000.5

"Flint equivalent" = 734

"Flint equivalent" = 730

AMERICAN CERAMIC SOCIETY.

109

The final ground coat is made by taking two parts of blue
ground coat and one part of white ground coat. This gives a
flint equivalent of 733.

Percentage composition.

Si02 67.82

A1203 3- 80

K20 1-77

Na20 6.66

B203 12.13

PbO 7 50

MgO 0.14

CoO 0.18

71 .62

100.00

Empirical Chemical Formula.
0.115K2O

0.646 Na20 { 6.827 Si02

0.021 MgO \ 0.224 A1203 {
0.203 PbO [ 1.033 B203

0.015 CoO

Ground Coat No. 6.

Batch for 1000 pounds.

Frit. Raw. Melted.

Sand 200. o

Feldspar 200 . o

Borax 390.0

Sodium nitrate 30.0

Red lead 100. o

Magnesium carbonate... 10. o

930.0 720.0

"Flint equivalent" = 732

Percentage
composition.

Si02 4800

AI2O3 l6.6l

64.61

K2O.

Na.O .

B2O3.

PbO..

MgO.

CoO.

Additions.

Clay 325 .0 279.0

Cobalt oxide. 1 .0 1 .0

1256.0 1000. o

3.38
7.42
14.27
9.80
0.48
o. 10

100.06

I I<>

JOURNAL OK THE

Empirical Ckrmical Formula.
0.167 M>

o.553 Na20 I (3-72 Si<)2

0.056 MgO [0.7.58 A1208 \
0.219 PbO [ 0.95 BjOj

0.005 CoO J

Ground Coat No. 7.

Batch for 1000 pounds.

Frit. No. 1.
Potash feldspar.

Borax

Sodium nitrate.
Red lead

Raw.

300.0

115. o

20.0

125.0

490.0

SiOj..
A1203.
KjO..
NaoO.
B203..
PbO..

61

Percentage
composition.

50.63

IO.7I

7. IO

4 36

6.95
20.30

Frit No. 2.

Flint 125 .O

Borax 75 o

Sodium nitrate 150

Red lead

Additions.

Clay

Flint

Feldspar

100.05

250.0

750

65

0

750

75

O

120.0

120

O

1 127.0 IOOO. o

'Flint equivalent" = 658

Empirical Chemical Formula.

0.31 K20
0.29 Na20
o . 40 PbO

0.428 A1203

[ 3.444 Si02

]

I 0.404 B20;,

AMERICAN CERAMIC SOCIETY. in

DISCUSSION.

Mr. Landrum: In the percentage composition given, do you
include the mill mix or is it merely the batch mix ?

Prof. Staley: That is a calculated percentage composition.

Mr. Landrum: But does it include the addition at the mill?

Prof. Staeey: Yes.

Mr. Landrum: And does the empirical formula also include
the mill addition?

Prof. Staley: Yes.

Mr. Landrum: Ground coat No. 7 is very similar to a sheet
steel ground coat.

Mr. Poste: I have been interested in some of the work Prof.
Staley has carried out in the past, relative to the possible substitu-
tion of one refractory for another and producing the same fusi-
bility. An interesting point is that this "flint equivalent" is ap-
parently the outgrowth of that work. I assume that the work
done heretofore was derived from actual experiment and has been
verified in practice. This is a much more complicated case of the
same thing. Is it practicable?

Prof. Staley: This was not tried out at all. These are for-
mulas which were obtained from various sources. I calculated
the "flint equivalent" from the factory formulas. In five for-
mulas, supposedly used for general sanitary work, bathtubs, etc.,
the "flint equivalent" was about the same and we had every
reason for assuming that about the same temperature is em-
ployed in most factories in maturing the ground coats. These
were not made up in order to verify the "flint equivalent." I
calculated the "flint equivalent" after the enamel had actually
been used in the factory.

Mr. Poste: What would be the effect of replacing the feld-
spar and other refractory materials by the theoretical amount of
flint? Have you determined this experimentally?

ii2 AMERICAN CICRAMIC SOCIETY.

Prof. Staley: Yes.

Mr. Poste: What were the results as compared to those oi
the previous experimental work?

Prof. Staley: I operated a factory for four years on that
basis, that is, substituting one material for another on the bash
of experimentally determined equivalents.

Mr. Poste: The old question of cobalt in the ground coat is
always interesting to us, and one little piece of experimental
work we were conducting a while ago brought forth the very in-
teresting proof of one theory; that under certain conditions in
steel ground coats containing cobalt oxide, metallic cobalt actually
separates and alloys with the steel. We happened to be doing
some fundamental work, involving mixtures which included nc
oxidizing materials whatever, but containing cobalt oxide. Aftei
withdrawing the stirring rod from the molten batches contain-
ing no oxidizing material, there was a peculiar glossy coating on
the surface of the rod. It appeared to be metallic cobalt. We
sawed off the end of this rod and found that metallic cobalt had
been deposited. We were endeavoring to determine the effect
of nitrates and we substituted increasing amounts of nitrates for
carbonates without varying the Na20 content, and the ten-
dency of the cobalt to deposit on the steel rod gradually disap-
peared, indicating that the oxidizing material kept it in the
oxidized form and in the absence of an oxidizing material it was
reduced, a deposit of metallic cobalt forming on the surface oi
the rod.

IE

'HE BONDING STRENGTHS OF A NUMBER OF CLAYS
BETWEEN NORMAL TEMPERATURE AND
RED HEAT.

By C. W. Saxe and O. S. Buckner.

The strength of clays at the temperatures and conditions to
/hich they are subjected prior to and during the early stages
if burning is of considerable interest, inasmuch as low strength
s often a direct and always an indirect cause of the breakages
idiieh occur between the time the ware is fashioned and the time
he burning operation has reached its maximum intensity.

In this connection the investigations1 by Kerr and Montgomery
lave shown that much depends on the manner and the thorough-
iess of the drying treatment to which a clay is subjected. Thus,
>wing to a small amount of water retained in its pores, St. Louis
ire clay, when in an air-dried condition, was found to attain
lot more than 50 per cent, of the strength that it possessed
vhen drying was completed at no° C.

Although the thoroughness of the drying treatment has a
narked effect on the strength, in factory practice it is not prac-
;icable, of course, to maintain dryers at temperatures high enough
:o remove all hygroscopic water. Even if it were, one or more
)perations on the ware are usually necessary after it comes from
he dryer, during which time it suffers a falling off in strength —
jwing to the absorption of moisture induced by room tempera-
tures and conditions. The loss due to breakage largely depends,
:herefore, on the selection of clays that are strong and tough
under the conditions of use.

In the most approved method of measuring the strength of
lays, the drying conditions are definitely specified and are con-
cluded with a temperature of no° C, which is continued until

1 Trans. Am. Ceram. Soc, 15, 270, 345 (1913).

ii4 I1 HTRNAL OF THE

constant weight is reached ' However, it is known thai km
clays when subjected to a drying treatment possess the property
of retaining moisture more tenaeionsly than others, and, on the
other hand, of absorbing it more readily when removed from a
dryer and allowed to eool at atmospheric conditions. Since the
presence of small amounts of water materially reduces strength,
it is probable that this affinity for water, shown in a variable
degree by different c'ays, has an appreciable influence on the
strength at conditions existing in the factory. If such lie the
case, then the relative strengths of a number of clays at factory
conditions would be expected to be somewhat different than their
order when properly and thoroughly dried at no° C

In investigating this matter, an attempt was made to deter-
mine: first, whether or not the strength of clay, dried in the ap-
proved manner at no° C, is indicative of the strength of clay
as it is likely to be used in the factory; second, to what extent a
number of representative clays in a dry condition differ as tc
their affinity for moisture; and third, the relative effects of mois-
ture on the strengths of these clays.

Another phase of this investigation is concerned with the
strength of clays during the early stages of burning. While
strength changes in clays in both the unburned and burned1
conditions have been shown, little has been published regarding
the changes in this property between i io° C and a red heat.

Tests therefore were made, first, to determine the changes
in hot and cold strength from normal temperatures to red heat
with particular attention to the changes undergone during the
dehydration period, and second, to note if the strength of clay,
thoroughly dried at a final temperature of no° C, could be used
as an index of the strength at the higher temperatures.

1 A satisfactory drying treatment and the one used by Kerr and Mo:
gomery is as follows:

a. Air drying for about seven days.

b. Drying to constant weight at 75 ° C.

c. Further drying to constant weight at no° C.

2 Bleininger, "Porosity and Strength of Clay Products," Trans. Am.
Ceram. Soc, 12, 564 (iqio>.

lit-

AMERICAN CERAMIC SOCIETY.

115

Experimental Work.
The trade names and general character of the clays used in
these tests are given in Table I :

Table i.

Clays Used in the Tests.1

Name M;irk

Kentucky No. 5 ball clay A Good plasticity.

Johnson-Porter No. 10 hall clay H

Armstrong clay C

W'hiteways hall clay D

Allen plastic clay E

Sehippach clay F

U. S. A. clay.

G

Michigan slip clay H

Albany slip clay I

Mitchell ball clay J

Good plasticity

Very good plasticity; not refractory.

Good plasticity

Good plasticity; a No. 3 fire clay.

Very good plasticity; German crucible
clay.

Very good plasticity; American cruci-
ble clay.

Poor plasticity.

Fair plasticity.

Very good plasticity.

For the first phase of this investigation each clay was tested
"or bonding strength, the tensile strength of a clay-alundum mix-
:ure being taken as a measure of this property. Each set of
:ensile strength briquettes was subjected to the following dry-
ng treatments:

(1) Dried in air, at 55 ° C, at no° C to constant weight and
:ooled in a desiccator over calcium chloride.

(2) Air-dried at room conditions.

(3) Dried in air, at 55 ° C to constant weight and then allowed
to stand at room conditions for 24 hours.

(4) Dried in air, at 55 ° C, at no0 C to constant weight and
then allowed to stand at room conditions for 24 hours.

In the first treatment, representing the method employed for
completely drying the briquettes, the second drying tempera-

1 Chemical Analyses of the Clays:

Mark. HjO. Loss. S1O2. Fe203. TiOj. AUOa. CaO. MgO. K2O

A 1.29 10.03 56.56 3.32 2.00 27.52

B 0.68 10.77 51.23 2.10 1.60 34.37

C 2.32 7.49 52.34 8.77 0.80 27.87

E 195 10.54 57 74 2.66 1.70 23.52 0.30 1.26 0.98 o

F 4.00 11.94 52.01 3. 1 1 o 75 31.99

H 4.42 6.33 60.66 2.83 1 . 12 12.28 7.77 2.55

I 1.21 7.69 55.73 4.94 1 .02 15 97 5.51 3.17

Total.
99-43

09

100

99
98

99
98

99

i [6 JOURNAL OF THE

lure, 550 C was used, but as the briquettes were largely composfl
of non-plastic material, this was found to give satisfactory I
suits. The remaining treatments approximate the averaJ
and extreme conditions likely to prevail in the factory.

In addition to the briquettes, trial peices of each clay were prt
pared and dried at the same time as the briquettes. The vvate
contents were then determined by drying to constant weight a
no0 C.

The results of this work are given in Tables 2 and 3 and Figs
1 and 2.

The strengths of the clays between normal temperatures an
red heat were determined on both hot and cold briquettes an
at such temperatures as to include the dehydration period. I
order to observe the commencement and duration of this period
loss of weight determinations and dehydration lag curves wer
made for each clay. Porosity determinations were also carriei
out perchance such data might be correlated with the bondinj
strengths. The results of these tests are given in Tables 4, 5, an
6, and are presented graphically in Figs. 3 to 13, inclusive.

Methods of Testing.

The Briquettes. — All of the briquettes were made from mis
tures of 25 per cent. 46-mesh clay and 75 per cent. No. 60 "regu
lar alundum" and molded into standard cement briquettes whil
in a soft plastic condition. Before testing, the faces of the bri
quettes were rubbed parallel and one end made flat in order ti
facilitate handling in the electric oven.

Drying Treatments. — The drying treatments for the first par
of this study were carried out as prescribed under "Experimenta
Work." Each operation at 55 ° C and no° C was continue*
till practically constant weight was reached.

Heat Treatments. — In the tests to determine the changes ii
strength during the early period of burning, the briquettes tha
were tested while hot at 200 ° C and above were held at the de
sired temperature for one hour befpre breaking, these tempera
tures being reached at the rate of 2.3 ° C per minute. Those tha
were broken cold were first partly cooled in the furnace itself anc

AMERICAN CERAMIC SOCIETY.

117

;n removed to a desiccator. Heat treatments below 200 ° C,
wever, were somewhat different in that they were continued
til the briquettes were constant in weight.
A Hoskins electric furnace was used for the treatments above
D° C, the temperature being measured up to 325 ° C by means
a mercurial thermometer and above this point by means of a
ermocouple.

Bonding Strength. — The testing machine used was a simple
vice and consisted chiefly of a lever of the first class fitted with
pair of clips or holders. The load was applied at the rate of

pounds per minute.
For breaking the briquettes in a hot condition, special refrac-
ry clips were prepared and connected with the arm of the lever
rough a small hole in the top of the Hoskins furnace. This
paratus is shown in the illustration.

Apparatus for breaking the hot briquettes.

1 1 s J( H K\.\l. OF THE

Six briquettes wire thought sufficient to provide a good aveJ
age result for the drying treatments al and belo\t [io° C. Pd
the higher heat treatments this number was reduce 1 to foil
and in a few cases three.

Loss in Weight. The trial pieces used for this lest were
•\ i" X 4" in size and were cut from a slip-made slab of clay whcr
in a leather hard condition. Raw, dry weights were obtained
after a final drying treatment at no° C to constant weight.
Bach test piece was held for one hour at the desired tempera-
ture, using the same rate cf heating as employed for the bri
quettes. When necessary, the bars were partly cooled in the
furnace and then removed to a desiccator. The per cent, loss
in weight is expressed in terms of the drv weights as obtained al
no° C.

Porosity was determined on the same pieces that were used
for loss of weight, by Purely 's formula,1

(W — D)

; X loo = per cent, porosity,

(W — S)

where

W = vSaturated weight,

D = Dry weight,

S = Saturated suspended weight.

(The saturating liquid was kerosene.)

Dehydration Lag Curves. — The dehydration curves were ob
tained by inserting a thermocouple into the center of a specimer
of clay 3 V' X 3/4" X i1// in size which was then heated at th<
rate of about 24 ° C per minute in an electric furnace having c
chamber 2" in diameter by 25/s" deep. Temperature reading;
were taken every 15 seconds. No rheostat was placed in series
with the furnace. The heating curve was obtained by using ar
already dehydrated specimen.

Results.

The calculated data given in Table 3 expresses the per cent
loss in strength that each clay has undergone on the absorptior

1 Illinois Geological Survey, Bull. 9, 142.

AMERICAN CKRAMIC SOCIKTV.

119

moisture, the latter being expressed in terms of the water
iuired by each clay to bring it to normal plasticity.

Discussion of Results.

3onding Strengths of the Clays. Relative Order of Strengths
Factory Conditions and When Completely Dried at a Final
mperature of no° C. Table 2, Fig. 1. — The bonding strengths
the clays when completely dried at a final temperature of
ftC are only roughly indicative of the strengths at factory
iditions. Not only is strength materially decreased by the
■sence or further absorption of moisture but the relative order
the clays, particularly the stronger ones, is changed. Thus,
y J, which has a bonding strength of 130 pounds per square
h when completely dried at a final temperature of no° C,
s among the class of mediocre clays when allowed to stand
room conditions for 24 hours. On the other hand, clay F,
ring a strength of only 88 pounds, is among the strongest
factory conditions. Clays C and G are further examples;
en dried to constant weight at a final temperature of no0

ISO

fle

HO

at 0

ifferent Origin

\ NcS Rail Claif
5 Ho 10 Ball Clay
' Armstrong Cla
1 Whitcwatj's Ba
'. Allen Plastic

a Treatments

I/O

t

100

j

1 Clat
C/ai/

1

0

<

r US A. Cloy

- Alt>oni/ 5lip Clay

la
«>$

*i
•i

|

s*

1. Mitchell Ball C/oq

A

f
1

1

cf

<4

B

a

1

t

70

0

Room Dry

SS'C Room
74 Hoars

IIO'C K>oom

2* Ho^rs

// o'c
Desiccator
Cooled

Fig. i.

JOURNAL OF THE

Tabus -•.

Bonding Strength in Pounds pbf Square Inch of the c'i.avs

AI'TKK Till ■; DffFERENT DRYING TREATMENTS.

Clay Air dried

A 32.9

B 30 >

C 65.3

1 » 49 7

E 58.9

F 64 1

G 526

1 46.1

J 57-5

Dried .11 5 5° ( Dried al 1 10° C, Dried at 1 10' C
Room 24 hours Room 24 boon De»i< < atoi

34 9
32 -5
69 9

52. I
59 .8
70.6
6l .2
51 -2
594

38.7

35 4
77 1
54 5
54-4
76.I
62 .0
57-3
63 -4

69 X
65 X

107.4

85.5

83 3
87.9

107 . 2
79 2

I30.5

Table 3.- — Per Cent. Loss of Strength per Per Cent. Moistlre Con-
tent Expressed in Terms of the Water of Plasticity.

Drying Treatment No. 4 at no° C, Room 24 Hours.

Clay.

A..

B..

C. .

D...

E..

F...

G...

I....

J...

A.
B.
C.
D.
E.
F.
G.
I..
J

Per cent.

loss in
strength.

Per cent.

water
content.

Water 0
plasticity

Ratio of the

per cent, loss

in strength

Ratio of to the ratio

the per cent, of the per-

water content cent, water

to the water content and

of plas- the water of

ticity. plasticity.

44 5

O.98

36.5

2.68

16.6

45

5

O.98

34

0

2.88

15.8

28

2

I .64

34

0

4.82

5 9

36

3

0.93

35

9

259

14.0

34

7

0.95

33

5

2.83

12.3

13

4

2 OO

33

8

5-92

2-3

42

1

I .21

35

4

3 42

12.3

27

6

0.55

22

2

2.48

hi

5i

4

I.24

39

2

3 16

16.2

Treatmen

t No. 3 Dried at 55 °

C, Room 24

Hours.

50.0

113

365

3 09

16.2

50

6

I-.I3

34

0

3

32

15 2

34

9

2 .02

34

0

5

94

5 9

39

0

1 33

35

9

3

70

10.5

28

2

1 . 22

33

5

3

64

7.8

19

7

2.89

33

8

8

55

23

42

9

1 79

35

4

5

07

8.5

35

3

0.63

22

2

2

84

12.4

54

5

1 95

39

2

4

97

10.9

AMERICAN CERAMIC SOCIETY.

121

they are practically identical but vary noticeably at the other
drying treatments.

Affinity for Water and Its Relative Effect. Table 3, Fig. 2. —
The data given in Table 3 show that there is considerable varia-
tion in the water content of the clays after the same drying treat-
ments. For instance, clay F absorbed from three to four times
as much moisture as clay I when dried at a final temperature
of no0 C and allowed to stand at room conditions for 24 hours.

The different clays, however, are very dissimilarly affected by
this moisture. For instance, clay J (Table 3), though it absorbed
less moisture than clay F, suffered a much greater loss of strength.
If the per cent, loss in strength per per cent, moisture content is
calculated for each clay it will be found to vary from approxi-

Pe

- cen

*■ Los.

in S

fren

?fh /

>er /

?er c

en+ i

ifthe

Wc

l+ert

f P/c

ishc

tyC

on fa

ned

in ft

e CU

7v

IIO'C

^

/

I* Hours

J

y

^

/ -

SS'C

Utouri

5

\

I

«fc

\

/

\

f

0

C lays
1 1 1

Fig. 2.

mately 7 to over 50 per cent. This great diversity in loss of
strength for equal amounts of water is still further shown by
Fig. 2. The results therein presented express the per cent, loss
in strength per unit amount of water of plasticity which each clay
retained or absorbed at two of the drying treatments. The
clays van- considerably — the smallest loss per unit amount of

I< M RNAL OF THE

water being approximately two per cent, and the largest over i6
per cent.

This greater susceptibility on the part of some clays to the eJ
fects of moisture seems to be the chief cause for the dissimilarity
of the relative order of the bonding strengths after the differenl
drying treatments.

Strength of the Briquettes Broken while Cold between Normal
Temperatures and Red Heat. Table 4, Figs. 3 to 13, Inclusive.
Aside from the increase in strength between normal temperatures
and 1 io° C, all of the clays show but small changes in strength
up to 3250 C. Above this temperature, and during the dehydra-
tion period, all increase in strength quite rapidly with the excep-
tion of the two slip clays, which remain practically constant
until red heat is reached.

Strength of the Briquettes Broken while Hot between Normal
Temperatures and Red Heat. Table 4, Figs. 3 to 13, Inclusive.
In all cases the strength of the briquettes when broken while
hot is lower than the strength of the briquettes similarly treated
but broken while cold up to and through the dehydration period.

Starting at 55 ° C, clay B shows the greatest loss in strength
and clay G the least, the former having decreased 20 per cent,
and the latter approximately 2.0 per cent. The average differ-
ence for all the clays is 10 per cent.

At no° C none of the clays show the characteristic increase
in strength that is evident in every case wrhen broken cold. In
fact, the tendency seems to be to decrease slightly below that at
550 C. The average loss of all the clays at this point compared
to the average cold strength at 1100 C is 28 per cent.

The majority of the clays at 2000 C show a further slight de-
crease, but at 325 ° C a slight increase is evident in most cases.
Above this temperature and through the dehydration period
strength continues to gain rapidly. The one exception is Mich-
igan slip clay wdiich remains practically constant up to 700 ° C.

Generally speaking, the clays that have the highest cold strength
at 550 ° C are also the strongest in the hot condition during the

AMERICAN CERAMIC SOCIETY.

123

Table 4. — Bonding Strength of the Clays between Normal Tem-
peratures and Red Heat.
(Pounds per Square Inch.)

Cold — desiccator cooled. Hot.

Room
dry
Clay. 20° C.

in —

A
B

C 89
D 61

E 85

F

G

II

I

J

38
36

86

73
14
57
61

60 62 76

59 66 73

114 134 129

80 94 98

107 125 121

91 114 127

104 145 145

19 24 17

75 88 80

89 125 no

84 126 163

78 129 158

133 151 468

103 151 266

119 184 322

141 253 331

153 255 294

19 9 24

87 92 274

in 195 221

O <N

49 52
47 49
98 98

74 73
94 96
88 84

102 88

17 15
70 66

75 73

53
57
89
61
84

7i
101

14
58

73

53

62

107

60

84

87

122

14
63
83

65

74

119

80

1 12

122

148

14

74

97

93
102

132
124

154
188
223
13
83
225

134 171
165 175
172 194

135 150
189 237
221 301
274 345

13 44
109 139
226 331

Table 5.-

Clay.

A

B

C

D

E

F

G

H

I

J

-Per Gent. Porosities after the Different Heat Treat-
ments.

no0 c.
34-9

37.
34
33
33
28

34
40
35
32

200°.
36.6
39-6
36.4
35-9
34-2
30.0
36.0
44.2
37-8
31-7

325

41C

°.

45 C

°.

57^

700°. 825

37-4

37-7

37-0

39-6

39-9 40-

40

3

40

7

39

3

4i

7

42

3 43-

38

3

38

O

37

4

38

5

38

9 39-

38

1

38

9

37

6

40

8

41

5 42.

37

7

38

5

37

4

40

3

41

2 41.

3i

O

28

2

30

4

33

1

33

2 33-

36

6

36

9

36

3

38

6

39

2 39

44

7

43

0

43

1

44

9

48

1 47

37

9

38

8

37

5

39

6

42

7 4i

32

3

32

4

3i

5

34

4

34

0 35

Table 6. — Per Cent. Loss in Weight at the Different Heat Treat-
ments.

Clay. 200° C

A o. 16

B 0.11

C 0.39

F> 0.34

E 0.52

F 1 . 19

G 0.29

H 0.25

1 0.09

J 0.36

325°.

410°.

450°.

57

5°.

700°.

8-\S°.

O.76

O.85

I. 1 3

7.90

8.09

8.69

O.80

O.66

I .32

9

36

9

55

I0.53

I.65

I .40

2.14

5

41

5

80

7.19

2.28

2.78

3-53

9

49

9

53

IO.90

3 24

3-43

4-55

8

72

9

03

9.84

1 .92

0.49

2-45

9

46

9

74

1 1 .20

1 . 10

1 . 12

1.86

10

05

10

61

II .60

0.28

0.36

0.63

2

74

7

14

7-55

o.35

0.33

0.81

3

05

7

09

7-59

0.97

O.83

1. 16

8

18

8

23

9.48

124

JOURNAL OF THE

<*

i

<

n

>.

«

«.

\ \

\

S

>

ll

5,

\

\

jt

- «. £ 1

-c- ^ 1;

\

, •»

\

•-•-.

-^.

t

«Q<&

"S\

T"\

!

"~"~

\

$

^ ^ ^ h§
? ^ Q

§

*

\ l

i

j

5i

1

\

$

\

1

?

V

i

§

; a

5

~t

\

\

k

I §

';

54

.g

~1 1
4$

I

'

g

'I
■3$

„

a

<fc

,

I!

-^
if

^1

\\

5

t

\

<

*

J

■^

\\

9<<

i

1

a.

\

\

"%,,

L

>

(j

!

<*:

1

~^\

\ 1

§S

|

c

|

\\

\

5 iS;

-j

\

V

?

■3

=0

\

I

5

1

S

,

I -

8 5

.

a

§

I

1

k

| k

1

"

£

t
i
<£

<S

AMERICAN CERAMIC SOCIETY.

125

1

>

5-1

s

*

5

1

„

^

\

1

6

^5

1!

1

\

\

-

- ^

-!

I

1

-T
5

v^>

"-.

^^

1

5^
- -5 »

\

^

>

>>

i,

I

" 1 "
. 1 _

\

\

I

I

2

1

1

1

\
\

g

J

\

\

j

? S

3 ^

Q

•a,

1

s

I

1

1

-?

^
*

<§

\

|

5

*,

9

»»

^

„,

,>

\

«a

1 ^-S $

\

» -

5- t

\

J

\

1

•

1

t

^

\

->

is

-5

1

f«1

it ■

3

\ \

1 £

§•

V

&

\ \

/

\

%

0

1

11

/

\

5

? s

§*

<a

a

§

1

1

H

0

^

i

1

fe

[26

l< fl RNAL OF THE

§

t

«

9

*

*>

l

■»

<-

\

i?

-1

1

-i.

|l

3l

\

\

•N

\

.

\

si".

1

4-'

<0

\

\

k-

•«.„,

>>

■s

**

^

1

"^

9

y I

• :

!

1

1

5.

\

' !

1

1

\

\

V

V

/

,

i

8

! R

a s

^

3
0

§

1

it

*t

1§

1

^

J

1

1

I

1

H

^

*

K

*

*

I*

1

'.

V

1

I

-I
.1

1
<!*

to "*■

!!

$ i

N

\

\

1

^

\

3*
^5

1

1

•w

V

1

I

v^

l\

§
§

8

^

■A

\

r

'J

s

n

1

1

^

^

1

^

1

r

!

\

'

\

l

/

\

\

■

n *

a §

<a

Si

1

1

I

^

( ?

'■

%

-?"

J

A

1

AMERICAN CERAMIC SOCIETY

127

\

^j

«,

s

k

5$

*

-

t 1

1

§

. >5

I

i

- 1
t

3

-i.
$1

' r

4

1

"I

u

■

^0
5.

g

g^

$

^

1

J

I

\

■v:
9 v

I

t
^

1

\t

1

i

1

/

i

\

\

s

s ^

a

5

§

u

§

k

{«

1

q

^

1

1

~S

s:k

5

5j

s

^

*>

> ^

*

\

\

;

s

|

J*

\

\

s-

\

*

\

ll.

\

1

,<3

L \

£

-^

.

j

*

1

^b

<*

"^\

\^

Qw

1

'--

\

ft

1

to K

$1

^>

}

1

i£

1

^
*

\

\

1

\

\

*

N

<s

s

* 5

S S

<&

m

§

Si

§

fc

'■;

^

^

i

1

1

128

JOURNAL OF THE

1

•*

«

5

*

■\

'.

x

K

\

§

■1

■1

I

.1

•v.

V

,\

\

1 1

£ s

Si

-1

\\

-SX,

r

1
1
1

$3

k

>

111

--,

--1

!-

.'4-

r

*S

S

\

\0>

^^

^~

? 5

*?

%

\
\

!i

1

\

\

/

\

\1

■'

S i

ft ^

*a

S

s

I

§

1

s*

1

<»

I

1

i

1

J

5

5j

5

1

n

M

^

X

\

8

>

\

\
\

-g —

-*>

- s;

■ 1

r
1

\

\

.

•^

t

\

x\

k

(

—4-

4

1 1

\

1

|i

a

\\

S

1 1

1

5

g

V s

S 5

^

§

1

i

^.

r

§

'-

•1

AMERICAN CERAMIC SOCIETY.

129

f

lehy

4ra +
A

on 1
B

oq

Cur
cot

'es

f 9

H

1

j

/

/

4

•V

2

/

>

/

HOO

/

/

0'

V

/

y

/

4

AS

/

1/

'A

/

u

<.y

1 1

/

j

/ J

1

<(W

/

f

/

^

§

A

/

A

//

/

/

/'

A NoS BallClatj F Scty,ppoch Clay
& NoiO Ball Clay 0 </ $ A Clay

C ArrrtstronqC/aij H tficn-qan Slip

0 wntteway's Ball 1 A/bony S//p
E Allen Plastic j fiitchell Ball

//

//

c 0

t

/

f

/
0

'

///

I J K

_|

K

Fior,da fSttolm

i 1

Fig. 13.

nterval between 55 ° C and 325 ° C, where strength is quite low.
AHien broken cold at the no° C heat treatment they are less
ndicative of strength during this interval.

Porosity and Loss in Weight. Tables 5 and 6, Figs. 3 to 13,

hclusive. — -No relation is noticeable between bonding strength
md the loss in weight of the clays at the different heat treat-
nents ; this property remains constant, increases or decreases as
he test pieces lose weight depending on the clay.

Likewise no relation is evident between porosity and strength
or the former increases gradually at all temperatures except at
)r just before dehydration, where, in every case, a slight de-
crease takes place.

Dehydration Lag Curves, Fig. 13. — The beginning of dehy-
iration was assumed to be at the point where the curve first de-
Darted from the normal furnace curve and the completion at the
joint where it first started to return. In every case the dura-
:ion of this period is somewhat briefer and not in very close agree-
ment with that indicated by the loss in weight determinations.
Hi is apparent discrepancy appears to be due to the errors attend -
ng the former method. Thus the absorption of heat, which ex-

i3o JOURNAL OF THE

tended in a number of cases beyond the upper limiting tempera
ture set by the loss in weight determinations, seems to have been
caused by the large size of the tesl specimens and the consequent
inability of the small electric furnace to regain its normal rate of
heating, following the endothermic reaction. The lag curves
for the two slip clays agree more closely with their loss in weight
determinations substantiating this view — for these clays are the
most impure of the lot and lose weight over a greater interval of
temperature.

Conclusions.

The strength of a clay, dried to constant weight at a final tem-
perature of no°C, is only a rough measure of its strength at con-
ditions which prevail in most factories. Results thus obtained
permit the selection of strong clays from weak clays but do not
afford a means of distinction between, say, a number of fairly
strong clays at conditions under which they are likely to be used.

When different clays are not thoroughly dried or are subjected
to atmospheric conditions where the absorption of moisture
is possible they show a large dissimilarity in their affinity for
wrater. The fact that the relative order of the strengths of clays
at factory conditions is appreciably different from their order
when dried to constant weight at a final temperature of no0 C
is partly due to this phenomenon. However, the largest and
most dissimilar changes are due to the fact that the presence of
small amounts of water causes some clays to lose much more
strength than others.

The strength of clays in a hot condition between no° C and
red heat is considerably lower for the first two or three hundred
degrees than their cold strength when thoroughly dried at a final
temperature of no° C.

During dehydration, both the hot and cold strength of clays
increases or remains practically constant indicating that the
physical and chemical changes during this period have little
effect on this property.

Generally speaking, the strongest clays in the cold condition
as dried at 55° C are in turn the strongest when hot during the
interval of lowest strength between no0 and 325 ° C and hence

AMERICAN CERAMIC SOCIETY. 131

the former can be taken as a rough measure of strength during
this interval.

COMMUNICATED DISCUSSIONS.

C. E. Fulton: This paper is very interesting throughout,
but the most important point which has been duly emphasized
by the authors is the fact that results obtained by tests upon
thoroughly dried briquettes (that is, dried at no° C and cooled
in a desiccator) will give results which will not be obtained under
practical conditions. It has been our practice, in testing the
bonding strength of raw clays, to air-dry the test bars for
one week, then place them in the drying oven at 75 ° C, and finally
at no° C until constant in weight. After this drying treat-
ment the bars are taken from the oven and, when cooled to room
temperature, are broken in a transverse testing machine. This
method has given very Consistent results and it will be interesting
to know whether or not the authors have any results obtained by
the breaking, as soon as they become cool, of briquettes which
have been dried at no° C.

With the exception of clay E, the results obtained with their
second drying treatment (55 ° C , in room 24 hrs.) and third dry-
ing treatment (no° C, in room 24 hrs.) give equally good classi-
fication. The clays are grouped in pairs, F and C showing the
highest strength, G and J forming the next pair, followed by D
and I, and finally A and B, which show the lowest strength.
The second testing method groups clay E with G and J while
the third drying treatment places it with D and I. As this is the
only clay which shows a falling off in strength between these
two methods, it is possible that the number of pieces averaged
to obtain this result (six) were too few to give it its proper posi-
tion. In all strength tests from ten to fifteen' briquettes should
be broken to obtain a fair average.

It would add to the value of the paper if data were given showing
the maximum variation from the average given for each clay, and
also a record of the room conditions during the 24-hour drying
period. This is one objection to the method proposed by the

i32 JOURNAL OF THE

authors as the room conditions introduce a variable which mal
make it impossible to compare results obtained at differed
periods and in different places.

M. F. Beecher: It is apparent that the drying methods
followed by the authors have not been made entirely clear. In
Table 2, the four columns represent values obtained as described
in the text under "Experimental Work."

The data presented in this paper confirm Mr. Fulton's state-
ment that an intermediate drying temperature between room
temperature and 1100 C. is required to give the best results.
In the present case, that intermediate temperature was 55 ° C,
The authors have no data to present on briquettes dried directly
at no0 C.

While in general practice it is agreed that at least ten briquettes
should be used to obtain an average, the large amount of work
of this nature which we have done has resulted in a method which
has been demonstrated to give very reliable figures even though
only six briquettes are used. The briquette mixtures contain
only 25 per cent of clay instead of 50 per cent as ordinarily used,
and granular, non-plastic of No. 60 mesh instead of No. 20 mesh.
This allows of more uniformity in the molding of the briquettes
and less danger of injury to the clay bond by checking in drying.
The machine and method of breaking, described in the paper,
contribute further to uniformity of results.

Some data on the accuracy of this method, compiled last year
for the use of the Committee on Standards of the American
Ceramic Society, show that in the test of 67 clays, only four
show an average deviation, from the mean tensile strength of the
six briquettes, of more than 7 per cent and that the grand average
of these values for "average deviation from the mean" was only
4.3 per cent. Reference to the work of Bleininger and Howat,1
shows a similar value for the tension method to be 6.1 per cent,
and the best result by any method was secured by the transverse
test where the corresponding grand average of deviations from
the mean was 5.5 per cent.

1 Trans. Am. Ceram. Soc, 16, 273 (1914).

AMERICAN CERAMIC SOCIETY. 133

Clay H, dried by the second and third methods described, shows
the following variations in the tension test:

Max. Deviation Min. Deviation Average Deviation

from Mean. Irom Mean. from Mean.

Per cent. Per cent. Per cent.

Method 2 9.0 1.5 5.3

Method 3 11.5 2.0 6.5

It was not the intention of the authors of this paper to propose
. new scheme for drying briquettes for bonding strength tests,
fheir object was to secure the necessary data for properly inter-
acting laboratory results on the strength of clays into usable
acton* information, and to point out that the difference between
aboratory and factory drying methods is such a great factor in
he drying of clays, that due consideration should always be
;iven this point in the application of laboratory values.

Incidentally, their data confirm the accuracy and reliability
>f the generally approved method of drying in the making of the
trength tests, to which Mr. Fulton refers.

Norton Co.,
Worcester, Mass

OBSERVATIONS ON THE FORMATION OF SEED IN
OPTICAL GLASS MELTS.1

By A. B. Williams, Pittsburgh, Pa.

In a number of optical glass melts, made at the Bureau of
Standard glass plant, considerable trouble was experienced from
seedy glass. In practically all cases the seeds formed were quite
large, varying from V32" to 1" in diameter, with a very few minute
seeds grouped about the outer edge of the areas in which the larger
seed were located.

No difficulty was experienced in the fineing of this glass, so far
as could be determined, the bubbles being completely eliminated
during the melting process and the glass remaining free from
bubbles during the working, so long as the temperature remained
above a certain minimum. These bubbles appear chiefly in our
crown and boro-silicate glasses, but have occurred to some ex-
tent in the barium glasses.

In some of the melts the bubbles have been formed in a layer
on the bottom, affording a cave-like appearance to the bottom
4 inches of the glass in the pot ; in other cases the bubbles extended
as a cone from the bottom, or the slowest cooled portion of the
pot, towards the center of the glass. In one case, the seed formed
in the center of the glass in the pot, the remainder of the glass
being free from bubbles.

The evidence which we have been able to gather in determining
the cause of this seed formation may be summed up as follows :

First, If the stirring of the glass were carried on below a cer-
tain temperature, bubbles would appear and increase in number
as the stirring was continued.

Second, If the stirring were discontinued at a temperature
above this minimum, and the pot were cooled without rapidly
chilling the bottom, a layer of seed, having a cone shape area

1 By permission of the Director, Bureau of Standards.

AMERICAN CERAMIC SOCIETY. 135

^tending upwards from 5 to 10 inches in the glass, would occur
n the bottom.

Third, If one end of the pot were raised off the bottom of the
lrnace during cooling, seed would appear in that portion of the
lass which was over the part of the pot still in contact with the
ottom of the furnace.

Fourth, If the cooling were started at a temperature above
lat at which the seed apparently first appeared during the stir-
ng and the glass in the pot was chilled rapidly, no seed appeared.

Fifth, The contents of the bubbles in samples of this glass
ere analyzed by Dr. Hillebrand, of the Bureau of Standards,
>r CO2 and small amounts were found present.

Sixth, Pots of these glasses which were not agitated but simply
lelted and fined in the usual manner until samples of the glass
ere free from seed, did not show any bubbles when cold. On
le other hand, there have been cases known in which pots of
>da lime glass, chilled by placing in the open, became full of
rge bubbles — this being contrary to the experience we have had
ith the crown and boro-silicate glasses.

It therefore seems very evident, from the above observa-
ons, that gases are evolved during the cooling of the glass at
imperatures at which the glass has still quite a low viscosity,
le temperature being below 12000 C. Slow cooling between
200 ° C. and 700 ° C. increases the amount and size of these
ed and rapid chilling prevents their formation.

In any case, we must recognize the fact that the formation of
ed in optical glass is not necessarily due to any difficulty in the
leing process, but, in many cases, it is due to some other phe-
)mena taking place during the cooling. The gas may either be
dissociation product or may be evolved from a supersaturated
■lution. Like phenomena occur in the cooling of steel ingots
id some of the silicates, such as diopside. However, the evolu-
:>n of gas in these cases is accompanied by crystallization and
e gas is probably evolved because of a decrease in solubility.

is improbable, however, that a decrease in solubility will take
ace upon the cooling of an amorphous glass melt, especially if
enry's Law applies to viscous liquids at high temperatures,

I>

JOURNAL OF THE

as it does to dilute solutions, ;md the k;is 's most probably a dis-
sociation product. The nature of this dissociation product tial
not been determined but as all of these glasses contain barium,
the possibility of the formation of Ba< h, which will dissociate to
BaO upon cooling, has been suggested.

Fig. i.

-Section of a pot of dense barium optical glass showing a mass of
bubbles in the center.

The accompanying illustration, Fig. i, is of a pot of dense
barium glass which was cooled from noo° C to 600 ° C in five
hours, the glass surrounding the light spot in the center being
free from gas bubbles and the light spot being a foam-like mass
of gas bubbles. This pot was cooled most rapidly from the top,
bottom and sides, the center having cooled more slowly.

DISCUSSION.

Mr. Gates: I would like to ask Mr. Williams if he has con-
sidered the possibility of a gas being introduced from the bottom
of the pot, causing the bubbles? I have seen cases of blistering
in a glaze evidenced by a similar effusion of gases, which evi-
dently came from the body of the ware, and, in cooling, formed
in the glaze. Could this gas, to which you refer, be derived from
the bottom of the glass pot during the cooling, as occurs in the
case of the blistering to which I referred ?

AMERICAN CERAMIC SOCIETY. 137

VIr. Williams? We at first thought that there might possi-

be some evolution of gases from the bottom of the glass pot,

I we have noted the formation of the gases in portions of the

ss which were not closely in contact with the bottom of the

VIr. Frink: I would like to ask if these pots are glazed
vious to being used?

vIr. Williams: The pots were glazed in the bottom only — be-
lting 100 pounds of cullet in the pot, previous to introducing
glass batch.

V[r. Frink: That being the case, it occurs to me that the
:stion which you have raised as to the increased solubility
jases in the pot itself, at its higher temperatures, would account
this condition. Perhaps the zone condition of the bubbles
also accounted for by the increased or diminished porosity
die pot or the solubility of the gases in those particular zones
±Le pot. If that is the case, when the pot is turned up on its
i, as we have often found to be the case, and is cooled
that side particularly, you will then have squeezed your gas,
ling up through the hotter portions of the pot, underneath
ir glass until it collects. The glaze, on the bottom of the pot,
ag more viscous on account of its higher content of alumina,
>ws the gas bubbles to collect until they become enlarged. In
king colored glasses, the pot is first filled and then, for some
son, the pot is cooled and taken out; you will find varying con-
ions in the size of these gas bubbles. The nature of the gas
>ends largely upon what the conditions have been and to
at period of melting the batch has been subjected.

Ar. Williams : I wish to state that our theory for the forma-
1 of the gas in the hotter end of the pot, when tipped up, was
t the slower cooling had taken place there, giving the gas a
.nee for evolution. The gases were always evolved while the
;osity of the glass was reasonably low and during the interval
:r which the slow cooling had taken place.

Ar. Frink: You do think though, Mr. Williams, do you
, that in view of the fact that the bubbles usually occur first

i j8 J( >URNAL OF THE

near t lit- clay surface, there is a possibility of these other gases beitJ
evolved from the pot itself and squeezed out and driven into the
glass?

Mr. Williams: I do not believe I do, because of the pressure
It would appear that any gas that was held in the clay pot mighl
be expelled from the outside of the pot rather than from the in
side. Another consideration would be the cooling down during
that period. If the gas were mechanically included, simply
present as gas in bubbles and held mechanically at the operating
pressure, there would be no expansion of the gases coming fron
the pot walls.

Mr. Frink: True, but I was proceeding upon the theory yof
have advanced, that the bottom of your pot, when first raised, ii
very hot, the central zone of the pot bottom being much colder
The central zone of the bottom that is in contact with the bench
must, if there is any difference, be the colder portion because il
is furthest removed from the source of heat. When the pot i:
raised on its edge and the bottom cooled, the glass next to the
bottom on the inside is hotter than the portion that is cooled 01
the outside; therefore, if there is a squeezing out, figuratively
speaking, of these gases, the gas must pass to the inside, the poi
being denser on the outside on account of the cooling.

Mr. Williams: I see your point, but our practice is to chil
the bottom with compressed air and by so doing, we do not have
this condition. W7e operate in this way believing that by se
doing the glass is cooled too quickly to allow the gases to come
out of solution. Gas bubbles do not appear when the pot i:
raised and quickly chilled. The bubbles are developed wher
the pot is allowed to set flat on the bench, the bottom remaining
hot for a longer time than the other portions of the pot.

COMMUNICATED DISCUSSIONS.

F. Gelstharp: This is a very interesting subject to a glass
maker, as the presence of seed and bubbles in the product is one
of his many troubles.

Bubbles of the large size to which Mr. Williams refers are neve]
found in glass except when the glass pots are allowed to coo!

AMERICAN CERAMIC SOCIETY. 139

m without special care being taken to prevent their formation.
■ experience agrees with that of Mr. Williams in that these
je bubbles are located in that portion of the metal which is
led most slowly. I have found that if the top surface of the
tal is kept warm and the bottom and sides are cooled quickly,
se large bubbles do not appear.

have examined these larger bubbles and in some glasses have
nd them to contain sulphur dioxide and carbon dioxide, and
)thers only carbon dioxide, but under very reduced pressure, so

in fact, that, when punctured under water they fill com-
tely, showing that the volume of gas is exceedingly small.
is would indicate that the bubbles originated from a very small
tntity of gas which has greatly expanded, owing to the surround-

conditions of greatly reduced pressure. To have this condi-
1 we must imagine the containing walls of glass to be at a tem-
ature of rigidity, that is the sides and the top surface of the
tal. As cooling of the whole mass progresses a condition of
atly reduced pressure will be produced in that portion of the
3S which is still mobile; this is generally near the center and
rards the bottom. Any small seed which may be in this locality
I be expanded to a size depending upon the viscosity of the
3S at that point.

am of the opinion that some of our optical glasses will evolve
solved or combined gases between certain temperatures, but
1 case of this kind the liberation of gas will be throughout the
Die mass, and the gas bubbles will not be very large in any
ticular portion of the pot, unless we have a condition of reduced
ssure as described above.

have seen old plate glass pots set out which had 9 to 1 2 inches
dass left in them and large bubbles of gas could be seen exuding
n certain spots, which were apparently cavities in the bottom
the pot. These bubbles would stay attached to the bottom

a considerable time till they were large enough to overcome

viscosity of the glass and rise towards the surface, but gen-
lly they were arrested before reaching the surface,
t seems to me that the bubbles found by Mr. Williams when

stirring of the glass was continued too long, may have been
I to the release of these bubbles exuding from the pot bottom —

i |«. AMERICAN CERAMIC SOCIETY.

the metal no1 being sufficiently fluid t<> allow them to come
the surface.

E. W. Washburn: Out experiments at Illinois have sho'
that pieces of perfectly clear glass entirely free from but!
contain considerable quantities of gas in solution. If ;i cli
piece of glass is heated until it softens and is then placed ii
vacuum the dissolved gases are evolved in such quantities
to cause the mass of glass to swell up to several times its origii
size and to assume the appearance of soap suds. Owing to 1
viscositv of melted glass it is not impossible that certain portk
of a pot of glass may be supersaturated with respect to the eq
librium between the dissolved gases and large bubbles of thi
gases. It is certainly true that all glass contains at le
that amount of dissolved gas which corresponds to compl
saturation at the maximum temperature attained during 1
process of fineing. A slight decrease of pressure at any ti
after cooling is started would be sufficient to start the evoluti
of this gas. Such a decrease in pressure might be brought abo
for example, by a variation of the firing conditions in the f
nace or by a fall in the barometer. Slow cooling would, of coui
give a greater time tor this result to occur, while rapid cooli
would tend to prevent it both by decreasing the time duri
which it was possible for such a fall in pressure to take place a
by increasing the viscosity to such a point that no growth of bt
bles was possible.

C. W. Keuffel: The phenomenon described by Mr. Willia
was observed in some of our melts, i. e., a cone of bubbles V
found to extend from the bottom of the pot toward the cent
This has happened to several meltings of Boro-silicate Crown l
never to a flint glass.

In each case we attributed this result to the rapid cooling
the top and sides of the pot and the relatively slow cooling of 1
bcttom. When the pot was allowed to cool either rapidly
slowly but uniformly from the sides, top and bottom no bubb
were formed.

Our Boro-silicate Crown contains no Barium so that the sugg
tion that the gas in the bubbles is a dissociation of Ba02 into B
seems in error.

AMERICAN CERAMIC SOCIETY
Summer Meeting.
Zanesville;, Ohio, June 25-28, 1918.

The meeting was an exceptional one in many ways and in
int of average attendance it materially surpassed that of the
rmer summer meetings. More than 90 were enrolled, seventy-
e of those in attendance having come from a distance.

lesday, June 2jth:

In the morning the members assembled at the Hotel Rogge,
.nesville, and after registration, proceeded to the plant of the
osaic Tile Company. The Dressier Tunnel kilns in operation
this plant were inspected, one being used in the bisque firing
d the other in the glost firing. Another interesting feature
this plant was the use of the Berg biick press as adapted to the
essing of wall tile.

In the afternoon the first plant visited was that of the Brush
cCoy Pottery Company, where the processes involved in the
sting and jiggering of pottery specialties proved interesting,
the J. B. Owens Tile Company, the structural beginnings of
Owen's tunnel kiln were observed and the details of construc-
>n and operation were explained. At the plant of the Ohio
ttery Company the manufacture of chemical porcelain was
ry interesting, this being one of the few plants in the country
oducing this line. The casting process is used for the most
rt in the forming of the ware and in burning, the European
ictice of a low temperature bisque and a high temperature
>st fire is followed.

In the evening a banquet was held in the dining room of the
)tel Rogge. An informal program, including songs and speeches,
ls afforded.

142 JOURNAL OF THE

Wednesday, June 26th:

In the morning the party visited the Roseville Pottery CM
pany, Zanesville. A tunnel kiln used in the burning of a
pottery and cooking ware was inspected. The party then pff
ceeded to the plant of the American Kncaustic Tiling Co. /
this plant particular interest was evidenced in the clay storai
bins and trolley systems, the use of compressed air in the si
presses, specially designed tile presses and the tunnel kilns en
ployed in the firing of glazed ware.

In the afternoon an excursion down the Muskingum Kivi
on the Steamer Louise was enjoyed.

On Wednesday evening the party proceeded to the plant 1
the Kearns Gorsuch Company, where great interest was man
fested in the processes of glass manufacture including the open
tion of automatic bottle blowing machines.

Thursday, June 27th:

In the morning the large paving brick plants of the Burtc
Townsend Co. were visited.

In the afternoon a motor trip was taken westward on tl
National Highway to the flint ridge district, Wilbur Stout, of tl
Ohio State Geological Survey, explained the ancient quart
workings where extensive operations were carried on original
by the moundbuilders and later by the American Indians, i!
one point the site of an ancient workshop in wrhich primith
implements w^ere made from the flint was visited. These flif
implements have been found as far west as the Rockies, as fc
east as the Coast and from the Lakes to the Gulf. A numbi
of specimens of these crude implements were found by membe:
of the party. The moundbuilders' fortifications at the Go
Grounds, Newark, were next visited.

Proceeding to Buckeye Lake a shore dinner was served t
the members at the resort hotel, after which a short busine
session was held.

Friday, June 28th:

The last day of the meeting was spent in visiting a variety <
industries in adjacent towns, the first stop being the Ransbo

AMERICAN CERAMIC SOCIETY. 143.

Bros. Plant in Roseville, (). Modern jar machines and
:r equipment used in the manufacture of stoneware were ex-
ned and demonstrated. At the Crooksville China Company's
it at Crooksville, a rotary dryer, which had replaced the mold
ier and old style pottery dryer, proved interesting.
1 the afternoon the plant of the A. E. Hull Pottery Co. was
ted and the methods of manufacturing white tableware
the single fire process was demonstrated. A unique method
lip decoration was a feature. From this plant the party pro-
led by train and trolley to New Lexington, Ohio, and visited
factory of the Ludowici Celadon Co. This factory is one of
most notable in the country and foremost in the manufac-

of roofing tile. The operation of a shale planer, cutting a
y foot face of shale, proved of unusual interest. The methods
aw clay storage and the preparation and feeding of the clay
his plant were exceptionally efficient. A producer gas fired
:inuous chamber kiln was explained. The methods of glazing

firing roofing tiles were also of special interest,
he last plant visited was that of the Ohio Brick and Stone
At the Company mine, flint fire clay is mined by the use of

mining machines. The clay is then passed through a wash-
plant to remove the iron impurities which occur in soft seams
he flint clay strata. In the manufacture of fire brick at this
it a portion of the flint clay is ground in a wet pan and used
i bond in the forming of the brick by the dust pressed process.

F. K. Pence, Chairman,
Committee on Summer Meeting.

THE HONOR* ROLL

OF THE

AMERICAN CERAMIC SOCIETY

Kenneth B. Ayer
H. H. Bartells
Charles E. Bates
Harold A. Best
Clarence J. Broekbank
John A. Chase
O. I, Chormann
H. F. Crew
Kenneth I. Fulton
Charles F. Geiger
Albert C. Gerber
Perry D. Helser
Herford Hope
Walter L. Howat
Thomas N. Horsley

* This list is not complete,
some branch of the service and
list will be appreciated.

F. Summer Hunt
Charles E. Jaequart
Joseph A. Martz
Fred A. Morgan
James S. MeCann
Edward Orton, Jr.
Walter S. Primley
C. E. Ramsden

G. W. Rathjeins
H. vS. Robertson
George M. Schaulin
Charles W. Thomas
E. H. Van Schoick
William G. Whitford

Information in regard to those who are in
whose names are not included in the above

C. F. Binns, Secretary, Alfred, N. Y.

IMS'

JOURNAL

OF THE

iMERICAN CERAMIC SOCIETY

A monthly journal devoted to the arts and sciences related to
e silicate industries.

g. 1 March, 1918 No. 3

EDITORIALS.

COOPERATION.

The need of technical as well as economic cooperation in the
;ramic industry is unquestioned. In the new era upon which
: are about to enter, cooperation will be the keynote of success.
More effective cooperation between our technical laboratories
d our manufacturers appears to be one of the needs of the
dustry. This is especially true in the case of the small manu-
rturing industries which do not have the services of a technically
lined staff at their disposal. Through the lack of technical
lidance there has been a great amount of wasted time and
ort in some instances. Laborious experiments and costly trials
iding to no results have and are being constantly made. The
:k of knowledge of some fundamental point has very often
elled the difference between success and failure. In the com-
tition which is destined to come after the War it will be neces-
ry to put our best into the industry. The methods of the past
11 no longer suffice.

In this connection it would appear that there has been a lack
cooperation and exchange of ideas among our technical labora-
ries. As evidenced by some of our literature of the past, there
.s been considerable duplication of researches and wasted effort,
several instances two laboratories have conducted parallel
searches covering the same ground. This should not be, as
r force of technically trained investigators is small and the field

•4'> H 'i RNAL OF THE

of Ceramic research is large. Through the activities of the cJ
mittees on Ceramics of the National Research Council we hJ

made an excellent beginning in the coordination of the wai
essential researches. Why not coordinate the peace-esserJ

researches through a central Research Committee as well?

HIGH TEMPERATURE FURNACES.

To our knowledge a satisfactory furnace for the determinate
of the softening or deformation points of fire clays and othc
refractories has not been placed upon the market. The ga
furnaces do not reach the temperatures necessary to the softenin
•of our most refractory products. The Deville furnace is unsatis
factory for several reasons. The granular carbon resistor furnace
have not come up to expectations. The vacuum furnace is ex
pensive both as to first cost and maintenance. The metho<
of determination of the softening points is being defined by ou
Committee on vStandards. It is to be hoped that a satisfactory
furnace for making the determinations will soon be found. In-
ventors please take notice.

INDUSTRIAL FURNACE SECTION.

The United States Fuel Administration has announced the com
plete organization of the Industrial Furnace Section, Bureau o
Conservation, to handle fuel conservation in all furnaces witl
the exception of those operated for the production of power, hea'
and light. This includes those plants using fuel for direct heat
of which the clay products industries are the most importan
branch.

Mr. A. F. Greaves-Walker has been placed in charge of thi
Section and has under his immediate supervision 13 district
comprised of 31 states, covering the territory in which industria
furnaces are used. Each district has a local head who has ii
his organization an advisory board and a corps of inspectin
engineers.

The newly appointed district chiefs met at the Fuel Adminis
tration Building, on July 29th, for a conference with the chie
of the section. The district chiefs made reports based on a pre

AMERICAN CERAMIC SOCIETY. 147

linary survey of each district. Their conservative estimates of
)bable annual savings were 3,000,000 tons of coal.
The clay products, lime, cement and glass industries will be
pected first. A standard questionnaire is being furnished each
nt owner in advance of the inspection, which he will fill out
i hold for the inspector. The rating of each plant will be based
mi the efficiency with which fuel is used. All wasteful burning
icesses will be taken into account, but an opportunity for cor-
ting wasteful conditions will be given before inspection. Reeom-
ndations will accompany each questionnaire, several items of
ich will apply to each and every plant, and if followed will
prove conditions materially. The inspecting engineer will be
a position to make further recommendations to each owner
er inspecting his plant.

iy appealing to the patriotism of the furnace owner, at the
rie time presenting tangible facts which show him a way of
'ing money for himself and fuel for the nation's use in the prose-
ion of the war, little trouble is expected in bringing about
lservation of fuel through the proper regulation and operation
all industrial furnaces.

MEMBERSHIP CAMPAIGN

Although we are passing through the vacation period during
ich it is always difficult to secure action — the Committee on
mbership reports the enrollment of an encouraging number of
v members during the summer months.

In active campaign for members will be carried on during the
and winter months and the organization of the industrial and
Sessional divisions as provided for in the Rules will be continued,
this connection it is interesting to note that formal action has
;n taken by the New Jersey Clay Workers Association leading
their affiliation with the American Ceramic Society.

ORIGINAL PAPERS AND DISCUSSIONS.

SOME PHENOMENA IN GLAZE REDUCTION.

By II. B, Henderson, Columbus, O.

In the development or finish of a number of wares, both dec]
live and utilitarian, reducing kiln atmospheres are import
factors. We ma}' refer to the employment of reducing fire;
the production of hard porcelains, in the darkening of salt gla
the change from dark brown to black, of red burning pa\
blocks and face bricks, and the golden flash on buff burning <
producK.

The production of these effects has become a common p:
tice and within reasonable limits, assuming the use of suits
materials, any desired effect may be produced by variation;
the kiln atmosphere and of the duration of the treatment.

We were confronted with this question in making laborat
salt glaze tests on samples of clay. Clays were submitted v
the request that we determine whether they were suitable for
production of salt glazed ware. Occasionally we secured fav<
ble results and could report the material satisfactory, but
tempts to duplicate the results on the same materials were n<
tive from which it was evident that negative results on a labi
ton- sample did not safely justify an unfavorable report,
difficulty was explained, however, when we observed that se
factor}' results were obtained and could be duplicated under
ceedingly smoky kiln conditions accompanied by a deposi
of carbon. It occurred to us that colloidal carbon depos
on and absorbed by the viscous surface of the piece might t
an important role in the resultant color of the glaze. We ki
that a very light yellow salt glaze has been obtained in a con

AMERICAN CERAMIC SOCIETY. 149

I kiln, under strongly oxidizing conditions, while wares from
same clay in a down draft kiln developed a dark brown glaze,
high temperatures, any kiln atmosphere, regardless of the
s oxygen content, has a somewhat reducing effect upon the
;s being burned. If the dark brown color developed by the
e in a down draft kiln is due to reducing conditions and
consequent presence of carbon, the light yellow color de-
ped by the glaze in a continuous kiln, under normal opera-
, is also due to a reducing effect, the carbon possibly being
essential factor in both instances, though the quantity of
on is not an important factor except in the development of
r.

he above statement represents the theory upon which we have
ned and carried out our investigations and the presenta-
of the subject at this time is merely a statement of progress
lout any attempt to draw final conclusions.

INVESTIGATION.

swer Pipe. — -Samples of salt glazed sewer pipe were collected
1 different factories, and microscopic slides were made from
nents of the glazes. Under the microscope, even at low
nifications, the glazes were found to contain crystalline ma-
1 and always of the same crystalline form, namely, hex-
lal plates. One would not expect any crystalline carbon to
ransparent and it does not seem reasonable that the crystals
graphitic although the carbon present may have influenced
crystallization of other constituents and affected the color,
lippings of the glazes from pieces of the Akron and Junction

sewer pipe were found to contain 0.01 percent, of carbon,
.mples of the clean glass drops taken from the roof of a kiln
le American Sewer Pipe Company, Uhrichsville, Ohio, were
d to contain 0.014, 0.012, 0.016 and 0.017 per cent, of
on, respectively. Under the microscope individual crys-
or several of them superimposed, could readily be studied
xansmitted light.

Fig. 1 is shown a microphotograph of a thin section made
l a dark, mottled and very brilliant glaze occurring on a No. 2

15°

JOURNAL OF THE

Tr v^'

^
t

AMERICAN CERAMIC SOCIETY. 151

clay body from Logan, Ohio — the dark spots being very
11 hexagonal plate-like crystals. In aggregates the crystals
ear to vary from brown to black in color, but when separated,
^appear amber colored by transmitted light. No particular
1 of grouping is noted.

1 Fig. 2 is shown a microphotograph of a reddish colored glaze
ped from a sewer pipe made at Junction City, Ohio, from a
ture of shale and fire clay. Several slides were made from

glaze, using fragments of the normal glaze and also frag-
ts from blotches of glaze which had dripped from the kiln
m. The latter fragments show an unusual growth of crys-
and in one instance, where the drippings had ■ collected on a
:et of the pipe, the appearance was very similar to that of an

oxide "aventurine" glaze. We made for comparison an
mturine" glaze (Fig. 3), and in thin sections the crystals
lis glaze appeared to be similar in form and color to those in
sewer pipe glaze, with the difference that the crystals in the
ter were larger, probably thicker and in consequence the
■ was more intense. In the "aventurine" glaze the crystals
mdoubtedly ferric oxide.
. Fig. 4 is shown a microphotograph of a glaze chipped from

2 clay sewer pipe made at Uhrichsville, Ohio. The blackening
iwer pipe by a strong reducing kiln atmosphere is quite a
mon practice and excessive reduction has a tendency to pro-
: crazing.1 An examination of samples of sewer pipe, differ-
n shade in consequence of variations in the kiln atmosphere,
rs an increasing number of crystals correlative to the dark-
ey in color. In case of crazing, it is possible that the glaze,
g heavily laden with crystals, is weakened probably by a
rence in the coefficient of contraction between the crystals
the glaze matrix and, in consequence, fine hair-like cracks

develop when the proportion of crystalline matter becomes
ssive.

Fig. 5 is shown a microphotograph of a fragment taken from
surface of a fire flashed brick. The golden colors of flashed
cs are developed under the same conditions of kiln atmos-

Larzen, Trans. Am. Ceram. Soc, 18, 434 (1916J.

t52 JOURNAL OF THE

phere as arc those of salt glazed sewer pipe. It" t he crys
present in salt glazes were the result of the formation of sodi
minerals, we would not expect to find them present in fire flas
products, in the development of which no salt is used. vSam]
representing various stages in the development of fire flas
bricks were prepared by scraping the surfaces with a hard s
tool, the dust thus obtained being mounted on slides.

Small particles of the crystallized material from the surf?
of the sewer pipe and flashed brick may be readily picked
and placed under the microscope by means of a needle p(
magnet. The magnetic properties of the crystals would ir
cate the presence of iron1 in considerable quantity but not ne<
sarily as a constituent of magnetite.

In order to illustrate the relative sizes of the crystals as she
in the microphotographs the following table of measurement:
given. The measurements represent in each case the aver
of a large number of determinations.

Sample No. i 0.00517 mm. .Sewer pipe, Logan, O.

2 0.00517 Sewer pipe, Junction City, C

3 0.01696 Aventurine glaze.

4 0.01680 Sewer pipe, Dennison, Ohio.

5 0.00266 Flashed brick.

Fig. 6 illustrates the variation in depth of color encounte
on the same product differing only in the degree of burni
"A" is a thin weak glaze; "B" is a glaze of good golden co
"C" is dark brown; and "D" is dark brown and crazed.

A determination of the number of crystals per unit area I
tive to the number in "A" as unity gave the following results

"A" = i, "B" = 6, "C" = 25, "D" = 20.

In D we find many imperfectly developed crystals, and perl
forms are seen entangled in a network of skeleton crystals. 1
condition is usually caused by very rapid crystallization.

Flashed Brick. — It has been generally maintained t
the color of fire flashed ware is due to a chemical cha

1 Trans. Am. Inst. Min. Engrs., Bull. 126.

AMERICAN CERAMIC SOCIETY. 153

le iron present in the product, but it is a well-

n fact that the color can only be produced by

ing the surfaces of the ware to contact with the

The face of a brick so exposed will be beautifully flashed

the ends or back not so exposed, yet not in contact with

bricks, and consequently exposed only to the general re-
g conditions of the kiln, will not be flashed. It is imprac-
to set the bricks in checkers because the flame in passing
gh the checkers will flash the surface under the open checker

the intermediate sections will not be flashed. The face of
Jc so exposed is barred with alternate flashed and unflashed
cs corresponding to the checker work. It is evidently
sary that the flame impinge direct on the surface of the
but it does not follow that any dust is deposited on the sur-

It might be that the local heating effect of the flame
mature develops a lower degree of viscosity, favorable to
rowth of crystals, caused to develop in the reducing atmos-
, and thus furnishing nuclei for their continued growth dur-
le subsequent period of oxidation in cooling. But we are
: opinion that the factor inducing crystallization is extraneous,
flashed brick when reburned in a normal kiln atmosphere
y loses its flashed color and comes back to the natural
:olor. It is a surmise whether the color of the crystalline
ial is eliminated during the oxidation or simply absorbed
e viscous matrix. We can conceive of the elimination of
n but it would seem that the iron would tend to give a more
>e color under the conditions of oxidation and solution fol-
g reduction.

ree samples of fire flashed bricks, all made from the same
■ial and the product of one factory, were selected for ex-
ition. One had a pinkish color, which is not desired; one
d fire flashed color; and one a yellowish buff color, or sim-

very light fire flashed effect. Under normal fire flashing
tions the pink color predominates but it is "burned out"
nidation before closing the kiln. In addition to the above
id a series drawn from the kiln as the cooling progressed,
ng the progression of the colors relative to the cooling.

i.S4 JOURNAL OF Till-

All of these samples show the same hexagonal plate cryj
found in the salt glaze. The tint of the pinkish colored al
is due to a few overgrown crystals of deep red color <:iit<
among the mass of smaller ones of uniform size and light ye
color. The larger the crystals the redder the color, the i
large ones approximating the color of the crystals in the"J
turine" glaze. The normal fire flashed sample contained a im
tude of light yellow hexagonal plate crystals of fairly unifi
size and color. The light fire flashed sample differs from
normal only in that there are fewer crystals. Trials drawn i
ing the cooling process show a progressive increase in the nun
of crystals.

Refiring under reducing conditions retained or developed
fire flashed color, but a well flashed brick, refired under oxidh
conditions, largely lost its golden flashed color and develo
instead a gray-buff color characteristic of a buff burning <
vitrified under oxidizing conditions.

An "old gold" flashed brick shows a mixture of crystal;
different sizes and colors, the larger approaching the red and
smaller having the yellow color.

Reducing conditions are necessary to the introduction or
velopment of the mineral constituents of the crystals, but
crystals are almost wholly developed under oxidizing coo
conditions, not that the oxidizing conditions are necessary — ;
haps quite the contrary. We know that in flashed brick the 1
temperature oxidizing atmosphere not only destroys the crys
but eliminates a factor or so changes the mineral constitui
from which h crystals form that we do not get a flashed surf
We also know, from practical experience, that a flashed effec
bricks can be developed at relatively low temperatures, but
effect lacks the true flashed color. Whether the color is du
iron or carbon the reducing conditions give a deposit on the
face of the ware — resulting in a dull lusterless flashed effect,
higher temperature there is surface fusion and in this fused r
the actual growth of the crystals takes place as the ware coo)

Briefly, in producing flashed brick we first subject the a
to reducing conditions, but if we continue this to the end, we

AMERICAN CERAMIC SOCIETY. 155

blacken the surface. Instead, however, after reducing we
•n out some of the black coloring element and start surface
ion. This is followed by additional reduction, partial burn-
out, the development of a greater degree of fusion, and thus
'mating to the end. As an alternative, we carry on the oxidiz-
conditions until the surface fusion is accomplished, then
uce and finally eliminate any excessive reduction effect by
elation, and finally cool slowly to develop the crystals,
f the color is due to iron, either in the clay itself or collected
11 the kiln gases, it is difficult to understand how this color
be burned out by subsequent oxidation except by the assump-
1 of transformation and diffusion effects.

t is inconceivable that, in the cooling stage when the color de-
jps, there can be any accumulation of iron from an outside
rce, but also inconceivable, were such accumulation possible,
t it could be dissipated by subsequent reheating under oxidiz-
conditions. It seems probable that the intense color is an
ical effect due to several modifications of crystals, or that the
I in solution in the crystal, is in a molecular condition, giving
laximum color effect which largely disappears when the crys-
ine iron compound is transformed or decomposed. That
eral modifications of iron oxide compounds occur according
the prevailing oxygen pressure in the atmosphere seems to
an established fact. Again, it is possible that absorption by
usion into the clay mass may occur.

Ahe color system as applied by Sosman and Hostetter1 indi-
es that the red and orange crystals contain an excess of ferric
ie while the yellow crystals have an increasing amount of
ous oxide.

Ve must conclude, either, that the fire flashed color and the
)r of the salt glaze is not due to iron alone or, that the peculiar
avior in its development and disappearance above noted is
: to physical changes. If the crystals are induced by carbon,
ir formation would be prevented under oxidizing kiln atmos-
res and they would not develop in the subsequent cooling — ■
ess the carbon be restored by an intermediate reducing kiln
dition.
1 Trans. Am. hist. Min. Engrs., Bull. 126.

[56 JOURNAL OF THE

Summary. Our preliminary work shows thai both the al
glaze and the surfaces of fire (lashed bricks contain erystallim
matter; that these crystals have a red to yellow color to whicl
the color of the producl is due and represent ferric oxide or othei
modifications of iron oxide. It is likely that the color of the- crys
tals is due both to a variable iron content, derived from the cla)
or the kiln gases, and modifications in the crystalline struetun
with variations in the temperature and oxygen pressure of the
atmosphere; that the salt glaze contains carbon; that reducing
conditions with the development of free carbon are essential foi
inducing crystallization; that crystallization hence starts in £
reducing atmosphere but continues under oxidizing cooling con
ditions; that re-oxidation largely dissipates the color effect ii
flashed clay products.

DISCUSSION.

Mr. ParmeleE: Is the firing under reducing conditions mon
favorable to the production of crystals?

Mr. Lovejoy: Mr. Henderson emphasizes the fact that th<
chippings from the glaze are magnetic, which indicates reducinj
conditions. He also quoted a statement made, I believe b?
Sosman, that the magnetic properties may be due to iroi
compounds other than ferrous oxide, to which the color may b
due. Reducing conditions are necessary.

Mr. Parmelee: May not the presence of metallic iron ac
count for the magnetic properties?

Mr. Lovejoy: No, it is an oxide.

Mr. Binns: Is it presumed that the color is due to colloid?
carbon or to iron crystals? Both were mentioned.

Mr. Lovejoy: The author states that the color is due to iro;
and that the part the carbon plays is not understood. Th
presence of the carbon appears essential, however, not as regard
the color, but as a factor in the development of the crystals. Tli
crystals probably owe their color to the presence of iron. Whe
flash brick are re-oxidized, the carbon present is probably elimrm

AMERICAN CERAMIC SOCIETY. 157

, and, in cooling, there is no development of crystals. If we
urn the brick under reducing conditions, we can restore the
pnal flashed effect. The introduction of the carbon seems to
the essential feature in the development of the crystals — ■
ither it is a chemical, physical, or mineralogieal phenomenon,
do not know. In the absence of the carbon the crystals
not developed.

£r. Parmelee: There is one point that is not at all clear.

have been shown that there is apparently a resemblance be-
en the crystals developed in the "aventurine" glaze and those
sewer pipe; is it true that reducing conditions are necessary

the development of "aventurine" glazes? If "aventurine"
ses are not developed under reducing conditions, the relation-
> between the two is not quite apparent.

1r. MinTon: I have produced a number of "aventurine"
:es and have never resorted to reducing conditions. I have
ays considered that oxidizing conditions were preferable. Mr.
rejoy emphasizes the importance of employing both reducing
. oxidizing conditions.

1r. Lovejoy: Mr. Henderson merely presented the "aventur-
' glaze crystal to show its similarity to the salt glaze crystal. He
s not state nor believe, so far as I know, that reducing condi-
ls are necessary to the production of "aventurine" glaze crys-
. In one salt glaze sample there was a glitter similar to that
an "aventurine" glaze which led to the comparison. He
s not discuss the "aventurine" glaze nor claim any relation-
» between it and the salt glaze crystal other than crystal form

color. When we first began to use natural gas in the
rking Valley, difficulty was experienced in securing the

glaze color that had been secured with coal. The desired
lit was later accomplished without the use of coal by increasing

volume and pressure of the gas and curtailing the supply
ir, in other words, filling the kiln full of unburned gases. My
ntion has recently been called to a practice which has been
eloped in connection with the burning of fire-flashed brick
1 coal. At the latter stage of the burn, a quart or more

[58 JOURNAL OF THE

of heavy oil is introduced into each furnace, the result beiil
very beautiful golden -flashed brick. The carbon undoubted
plays some part in the development of these colors.

Mr. Washburn: Attention has been called to the prcsei
of carbon as a necessary condition in the development of I
color. The effect of the carbon is to produce a certain defin
concentration of carbon monoxide, and it is in this indirect mani
that the reducing action takes place. The carbon itself does 1
appear to play any direct part in the action, but insures a defin
concentration of carbon monoxide which apparently is the act
reducing agent.

Mr. Binns: Some of those here can recall, in connection wi
the work of this Society, about 18 years ago, a discussion in ref
ence to the development of flashed brick, and one of the me
bers displayed a series of trials, drawn at periods from a ki
showing that the trials when first drawn were blue, but the co
gradually changed to yellow in cooling. Founded upon tl
fact, a theory was advanced which has been more or less the si
ject of discussion ever since — the formation of ferrous silicat
The thought which has been passing through my mind in that a
nection, applying to both this case and that of flashed brick,
that ferrous silicate has evidently formed and may be oxidii
to the ferric silicate. If this is true, it accounts for the phenom*
of flashed brick and also the yellow crystals in a carbonacei
magma. I think that this point explains the difference betwi
the "aventurine" glaze and the crystal in the sewer pipe,
the "aventurine" glaze, the iron is added to the glaze anc
readily available; in the sewer pipe the iron has to be dissol
from the clay and can only be dissolved when it is in the fen
form. I have produced successful salt glazes in an experime:
kiln without any reduction and have secured brown colors qj
easily with no more reduction than is caused by a partial clo:
of the damper.

Mr. Stull: Certain shades of amber glass are produced
dissolved carbon and the presence of a small per cent, of car
in the glaze would account for the amber color. We know t

AMERICAN CERAMIC SOCIETY. 159

1, under certain conditions, will also produce a yellow color.
is possible that the yellow color is enhanced by the presence
the dissolved carbon.

Ar. Binns: Why is it then, that in the burning of porcelain,
carbon stains the glaze black? The greatest difficulty in

ning porcelain successfully, under reducing conditions, is the

lination of the carbon dissolved in the glaze. Why is it that
carbon stains the sewer pipe glaze yellow, and the porcelain

it black?

/Ir. Stale y: In regard to the formation of ferric silicates, I
ild say that one encounters statements in ceramic literature
he effect that ferric silicates cannot be formed and that ferrous
:ates are readily formed. We have, in mineralogical litera-
t, repeated references to ferric silicates. In acid glazes one
introduce red ferric oxide, which, upon heating in the elec-
furnace, produces a yellow glass. There is no question
hat case of the formation of a ferric silicate, without previous
nation of ferrous silicate.

Ir. Binns: Mr. Staley has no proof that it is not a ferric
ition.

NATURAL HYDRAULIC CEMENTS IN NEW YORK

BY R.OBSRT W. JonKS, Tompkinsville, N. Y.

It is impossible to determine the exact date when natural
rock cements first came into use. In modern times the first
structure, of any importance, in which this material was used
was the Eddystone Light House, constructed by John Smeaton.
The original discovery by Smeaton was in 1756. It was not until
1796, however, that the actual production of natural cements
was made on a commercial scale. This material, produced from
septaria, was patented by a Mr. Parker, under the name of
"Roman" cement. Twenty-two years later the first production
was made in the United States. The growth of the industry, in
this country, has been closely associated with the construction
of canals and other improvements of our inland water ways.

Limestones, after being burned, produce materials which can^
be conveniently classified under four heads, resulting chiefly
from the varying amounts of impurities in the form of silica, iron
and alumina. This classification includes common lime pro-
duced from the comparatively pure calcium or calcium -magnesium
limestones; weak hydraulic limes carrying from ten to twelve
per cent, of impurities ; strong hydraulic limes with from eighteen
to twenty-two per cent, of impurities, and hydraulic cement carry-
ing as high as forty-two per cent. A large production of material
reported from New York State as hydraulic cement should have
been classified under hydraulic limes. The production of hy-
draulic limes, as a commercial industry of any magnitude, ceased
many years ago. The following figures of production have been
•corrected, as far as possible, in regard to this production.

In 1768 was the first official notice of the desire for improv
ment of the inland navigation of the present State of New York.
It was not until 1791, however, that any official act was passed
looking towards the fulfillment of these desires. In that year
there was appropriated by the State Legislature the sum of $250
for an exploration and survey along the line of the present Barge

AMERICAN CERAMIC SOCIETY. 161

Janal. This work was completed under the direction of Benja-
nin Wright. In 1816 Mr. Wright was assigned as engineer in
charge of the middle section of the Erie Canal with jurisdiction
extending from the Seneca river to Rome, N. Y. Mr. Wright
lad as assistant engineer Mr. Canvass White. Mr. White, who
was an engineer of great industry, made a voyage to Europe, at
lis own expense, to study canal construction. While in Europe
:he use of Taris and Roman cements, in hydraulic construction,
vas strongly brought to his attention. Mr. Benjamin Wright
lad recommended the importation of these cements for use on
:he Erie canal, but the great expense attached to this importa-
:ion had worked against it and attempts were being made to
complete the rock work with lime mortar. In most cases this
vas attended with complete failure.

In 1 818 Mason Harris and Thomas Livingston entered into a
contract to furnish lime for the middle section of the Erie canal.
\. quarry was opened on the property of T. Clark, Chittenangc,
Viadison County. The burned material from this quarry lacked
;he usual characteristics of the ordinary lime in that it did not
ilake. Samples of this material being brought to the attention
)f Mr. White were recognized as being similar to the English
R.oman cements with which he was so familiar. After consid-
erable experimentation patents were issued to Mr. White on
February 1, 1820, and February 17, ,1821, for the manufac-
;ure of natural cements for hydraulic purposes. There was con-
;iderable reluctance on the part of the Canal Commissioners and
nore on the part of the masons towards the use of this hydraulic
cement. After a number of failures in the use of lime mortar,
lowever, it came into universal use. In 1825 Mr. White brought
;uit against several contractors for non-payment of royalty,
imounting to $0.04 per bushel, and secured judgment. The
contractors petitioned the legislature to relieve them of these
udgments. The joint committee on Canals and Internal Im-
)rovements, believing in the justice of Mr. White's claim, recom-
nended that he be paid the sum of $10,000 to secure, to the peo-
)le of the State, the free use of his cement. Mr. White accepted
md a bill was introduced on February 11, 1825, read the first
ind second time, and committed to the committee of the whole

r62 JOURNAL OB THE

House. The bill died there and no further attempt was made
to secure the payment.

In 1823 Mr. Wright was instructed to make a canal survey
from tidewater, at the mouth of the Walkill, in Ulster County,
to the Delaware river. This was to be known later as the Dela-
ware and Hudson Canal. Construction began in 1825. Chit-
tenango. cement was used in the rock work during that year.
Later in that year hydraulic cement rock was discovered near
High Kails, Ulster County, but it was not until the summer of
1826 that the first commercial kiln was burned. This was at
High Falls, and was operated by John Littlejohn. From this
came the great industry which made Rosendale cement known,
throughout the United States, as the standard natural hydraulic
cement. After the completion of the rock work along this canal
the lecal cement industry declined for a number of years and, un-
til 1847, was carried on in a small way by a great number of pro-
ducers. In that year came the first combination of operators
and the production increased rapidly from an annual output
of 200,000 barrels until in 1898 there was produced, from this
section, almost 4,000,000 barrels.

For many years, after the discovery of the natural cement
rock at Chittenango, there was great activity in the natural
cement industry. Rock suitable for the production of this ma-
terial was found at many localities along or in the neighbor-
hood of the Erie canal and kilns were soon in operation through-
out nearly the entire length of the canal. Soon after construc-
tion was begun on the Champlain canal, cement rock was dis-
covered at Galesville, and quite a production was made for use
on this canal. The construction of the Black River canal was
responsible for the discovery and production of natural cement
at Lowville, Martinsburgh, Depeauville and Waddington. This
section supported quite an industry, although it declined rapidly
w'ith the close of canal construction. The more prominent locali-
ties along the Erie canal were Grand Island, Williamsville, Buffalo,
Akron, Lockport, Oak Orchard Creek, Chittenango and the
localities in and around Onondaga County. These were Fayette-
ville, Manlius, Edwards Falls, Jamesville, Marcellus Falls and
Skaneateles Falls. Fayetteville began to produce in 1820 and

AMERICAN CERAMIC SOCIETY.

163

Williamsville in 1824. In 1839 Jonathan Delano erected works
at Falkirk, near Akron, Erie County. In 1843 this business was
passed on to James Montgomery and later transferred to Enos
Newman. In 1854, H. Cummins and Son erected a plant at
Akron. This was succeeded in 1865 by another plant. The
Akron Cement Company was organized in 1870, taking over the
plants of H. Cummins and Son. A few years later Cummins
Brothers erected another plant west of Akron. The production

Fig. i. — Behan's quarry, Manlius, Onondaga County, New York. The
upper level of the quarry floor and the second heavy bed from the top of the
quarry face were the productive cement layers.

of cement was carried on at Akron until 1909. This section was
probably the greatest producer of hydraulic limes in the State.
The first cement made in Buffalo was produced by Warren
Granger, in 1850. Lewis J. Bennett began the production of cement
and organized the Buffalo Cement Company in 1874. This was an
active producer of cement until 1902. The plants in Erie County
are now all abandoned. The only other section, along the Erie
■canal, to become a permanent producer, was the Onondaga

"' \

JOURNAL OF THE

County section. Production began ;it these localities almost
immediately after the original discovery and has been eon
tinuous to the present time. The output of this section
falls in a elass between the true hydraulic cements and
the hydraulic limes. It can hardly be classed as a true cement,
although its characteristics place it closer to the cements than
the limes.

The only other section making a continuous production was
that of Howe's Cave, Schoharie Countv. The first kiln was

Fig. 2. — A continuous type of cement kiln, Cummings Cement Co., Akron,
Erie County, New York. This was the type of kiln used by all of the Erie
County producers.

burned at this locality by Mr. Vrooman in 1867. A few years
1 ater another plant was placed in operation and the two were
productive until 1896, when they were consolidated and, after
considerable reconstruction, production was again started in
1898, by the Helderberg Cement Company and carried on until
1905. At the present time, in the State of New York, there is
one plant in operation in the Rosendale District and three in
Onondaga County. The indications are that only one company

AMERICAN CERAMIC SOCIETY.

165

rill celebrate the one hundredth anniversary of the discovery
►1 natural cement in this State.

Geology. — The production of natural hydraulic cement has not
>een limited to any one formation, but includes limestones from the
niddle Ordovicic to the upper Siluric. The lower, or oldest in
he geological scale, is found in the Black River beds. The pro-
iuction from these beds was short lived and was consumed al-
nost entirely in local construction. This material is included
imong the true cements, although of a rather poor grade. These

ig. 3. — Opening in the cement layers at Akron, Erie County,
layer of Onondaga limestone forms the roof.

A heavy

)eds have been productive to the extent of approximately 250,000
barrels. Next in ascending order comes the Rosendale water-lime,
1 subdivision of the Cayugan group of the vSiluric. It is from
:hese beds at High Falls that the original discovery, in the Rosen-
lale region, was made and has been responsible for a production
greater than any other section. Next in order come the Rondout
vater-limes of the Cayugan group. Rondout water-limes are
exposed in the region between h)ast Kingston and Kingston
ind have been extensively worked. The production from Buffalo,

[66

Jl >1 RNAL OF THE

Akron and Howe's Cave came from this horizon. The formation
highest in the geological scale, which has furnished a commercial

natural cement, is that of the Manlins limestone. Near the top
of this formation, in Onondaga County, are found two thin beds
from which the entire output of this section has been produced.

Mining. -In ( hiondaga County, due to the few feet of material
available, the production of crude rock has been secured from open
cuts and presents no features of unusual interest as regards min-

Fig. 4. — Openings of cement mines at Rosendale, showing method of mining
along the productive beds.

ing. In the Buffalo-Akron district three of the four producers
secured their crude material through underground mining. Adits
were driven in along the outcrop, leaving heavy pillars of cement
rock for support. The cement rock, varied in thickness from
five to eight feet. The roof, being of massive limestone, re-
quired only a very small amount of timbering. The Howe's
Cave material was secured in the early days by open cuts and
.later by underground mining. It was in the Rosendale-Rondout

AMERICAN CERAMIC SOCIETY. 167

region that the greatest difficulties were encountered in the pro-
duction of crude material. In this region the formations have
been folded and faulted to such an extent that the entire output
has had to be raised through inclines to the kilns. A distance of
over eight hundred feet along the incline has been reached in one
instance. The mine of the Consolidated Rosendale Cement Com-
pany, at Binnewater, is a typical example of mining in this region.
At this property the main incline has an inclination varying from
thirty-five to fifty -five degrees. There has been no attempt
made to place the different working levels at any regular inter-
val although the distance does approximate one hundred feet.
On arriving at a point on the incline where it has been determined
to drift along the beds, a cross-cut is driven across the three
workable beds, which are known locally as the "old dark," "new
dark" and "light." Between the "old dark" and the "new dark"
beds is another bed of "light" which, on account of poor mining
conditions, is not productive. Drifting is carried along the
three beds in a northerly and southerly direction and at approxi-
mately right angles with the cross-cut. The slight grade of the
drifts towards the cross-cut throws the drift line on a slight curve
towards the east in order to keep out of the foot wall. At a dis-
tance of about sixty feet north and south from the cross-cut,
along the "old dark" bed, inclines are driven downward along
the bed. On reaching the determined new level, drifts are run
connecting the two inclines and a raise is then driven along the
bed in line with the main haulage incline. Rails are laid as the
work progresses and, preferably at the end of a Saturday day
shift, connection is made. On Monday the new level and in-
cline is ready for use. At each level the opening of the haulage
incline is equipped with a hinged door so arranged that when
lowered the descending car passes over into the cross-cut and
when raised allows the car to pass to the lower levels. There
are six levels, the lower being used as a sump, the two upper
not being productive on account of poor hanging walls. Very
little timber is used and only at points where the hanging wall
shows evidences of weakness. The working faces are carried with
a width of about twenty -five feet and a height of eighteen feet.
The entire output comes from the drift faoe and no attempt is

[68

JOURNAL OF THE

made towards stoping. Operations on the first and second levels
had, in the early days, been let by contract with very poor re-
sults. The first Level has been carried to the surface by frequent
raises along the bed with very small pillars placed at irregular
intervals and leaving very large rooms. It had been the inten-
tion of the present Company to mine the lower levels, leaving
pillars of not less than forty feet in width.

Burning. — In the early days the raw material was broken roughly
by hand and burned in crude upright kilns similar to those used in

Fig. 5. — Interior view of mine at Rosendale, Ulster County, New York.

burning lime. The walls were usually built of limestone and the
kilns were sometimes lined with brick. Wood at first was the
only fuel economically available. The fuel consumption was
high and as soon as coal became more plentiful it was substituted
for the wood. With the use of coal came the permanent upright
kiln into which the crude rock and coal were fed in alternate
layers and a certain amount of burned material removed every
day. Rotary kilns were never used. The burned material was
hand-picked, the underburned being fed back to the kilns or to a

AMERICAN CERAMIC SOCIETY.

169

special kiln and the burned being passed through jaw crushers
and finally to millstones. After ageing, the material was shipped
in paper sacks or wooden barrels.

Fig. 6.1 — Alvord and Company, Manlius, Onondaga County, New York.
Old style of intermittent kiln now abandoned except in this district.

The following figures of production have been taken in most
cases from original sources and may be accepted as the correct
number of barrels of natural hydraulic cement manufactured in
the State of New York since the discovery in Madison County.
No attempt has been made to calculate the value per barrel.
The production by districts is as follows:

Rosendale-Rondout district 85,873,000 Barrels

Erie County 14,300,000 "

Onondaga District 17,000,000 "

Howe's Cave 1 ,250,000 "

Miscellaneous 900,000

Total 119,323,000 "

1 The preceding illustrations have been taken from cuts which have ap-
peared in various reports of the New York State Museum.

TESTS OF CLAYS AND LIMES BY THE BUREAU OF
STANDARDS PLASTICIMETER.1

Hy F. A. Kirkpatkick and W. B. Orangk, Pittsburgh, I'a.

1 By permission of the Director, Bureau of Standards.

It is the purpose of this discussion to give the recent develop-
ments in the theory and use of the instrument described by W.
E. Emley in the Transactions of this Society,2 as the "Bureau of
Standards Plasticimeter" as follows:

2 Trans. Am. Ceram. Soc, 19, 523-33 (1917):

"When spreading a plaster on a wall, the trowel is so held that it makes
an angle of 10° to 15 ° with the plane of the wall. The force applied to the
trowel is thereby divided into two components, one acting parallel to the
wall — the other acting normal to it.

"In the plasticimeter, the trowel is represented by a conical disk, which
is mounted point downwards on a vertical shaft. The side of the cone makes
an angle of 10° with the horizontal corresponding to the angle between the
trowel and the wall.

"The wall is represented by a disk of plaster of Paris of known absorption.
It is mounted on the upper end of a vertical shaft, directly below and con-
centric with the cone. This shaft is threaded and runs through a fixed nut.
As the shaft revolves, the wall (or disk) is forced against the trowel for cone)
with a spiral motion. It presses vertically upwards against the cone in a
manner analogous to that component of the force exerted by the trowel which
acts normal to the wall. The turning motion moves the disk over the face
of the cone bringing into play that component of the force which is parallel
to the wall.

"The cone is so mounted that either of these forces is able to cause motion
independently of the other. Any motion of the cone is transmitted through a
system of levers to one of two bars, which are suspended like pendulums,
and causes the bar to swing out of its originally vertical position. The
greater the angle through which the bar is moved, the greater will be the force
of gravity tending to return it and the cone to their original position. It
follows, therefore, that if a constant force tends to cause motion of the cone,
the bar will swing out until the force of gravity acting on the bar is in equilib-
rium with the force acting on the cone. When this condition has been at-
tained, both the cone and the bar become stationary, and the force acting on
the cone is directly proportional to the sine of the angle through which the bar
has moved. The tars are graduated and provided with bobs which can be
moved up or down thus decreasing or increasing the force required to produce
a given angular deflection.

AMERICAN CERAMIC SOCIETY.

i?i

•'Calculating from the dimensions of the instrument as at present de-
signed, we find the following relations:

v = 1.6 + 27.42 sine a + 4.65 1 sine a.

/ = 9.29 sine a + 1-577 J sine a-

v = vertical force acting on cone, in grams per so., cm.

/ = tangential force acting on cone, in grams per sq. cm.

a = angle which the bar makes with the vertical.

/ = distance from point of support of bar to top of bob, in cms.

Q) ;) BC/tfEAU OF STANDARDS /=>LAST/C/M£TER

piAsret? block

0

■fSv

172 JOT RNAL OF THE

"The independent term in the lirst equation is caused by the weight of
the com- itself

"If a sample- of plastci is molded on the disk, and is driven against the
rone at a constant speed, the two l>ars will assume definite position Kiom

these positions, the two forces acting on the sample may be calculated by

means of the above formula. The vertical force corresponds to that com
ponent of the force actuating a trowel, which is normal to the wall, tin
tangential force corresponds to the component parallel to the wall. Thi
machine is, therefore-, capable of measuring independently the dill'en nt
forces which are applied to a trowel."

In the use of lime the word "plastieity" may be used to indi< ate
the intrinsic plasticity of a paste which is not losing water to an
absorbent surface ; the rate of change of plasticity of a paste spread
upon an absorbent surface; or the degree of workability of lime
mortars. In each of these three cases different laws apply to the
observed behavior of the materials. Hence, until more work has
been done with the plasticimeter and the word "plasticity" is
defined for the various materials, it is not advisable to offer the
instrument as a universal means of measuring plasticity.

In the case of clays, it has not been determined whether or
not the plasticity is actually measured by this instrument. The
experiments with clays in the plasticimeter have been based upon
the changes in plasticity when in contact with an absorbent
surface. It was hoped that by the use of this instrument the
relative values of clays and pottery bodies in the process of
jiggering could be measured.

The authors of the present paper do not agree with Mr. Emley's
calibration of the instrument so far as the tangential forces are
concerned. In making this calibration he assumed that the
forces acting in the horizontal plane, between the disk and the
specimen, were of uniform magnitude in each unit area of the
disk surface and hence that the total force could be considered as
concentrated at a distance from the center equal to the radius
of gyration of the surface of the disk. It is well known that the
forces of friction vary with the speed. The speed of different
portions of the specimen varies as the distance from the center
and the forces of friction and adhesion are also subject to varia-
tion. The laws of these variations for lime pastes are not known

AMERICAN CERAMIC SOCIETY. 173

and considerable work will be required to determine them.

In view of these facts we have not used the equation given in
Emley's article: In the present work the distance / was 12 cm.
and was not varied. The sine of the angle a was used as "the
relative tangential force." This method gave numbers which
were proportional to the actual forces involved and eliminated
the uncertainty of the calibration originally proposed. A mea-
surement of the vertical forces has been found to serve no useful
purpose.

Valuable results have not been obtained for either clays or
limes by the use of a non-absorbent surface. In the tests herein
reported a plaster of paris absorbent surface was used. As will
be shown, contrary to Mr. Emley's contention, the results were
found to depend upon the consistency of the clay body or the
lime paste.

The method for determining the sand carrying capacity of the
lime, as proposed by Mr. Emley, was found to be impracticable.
Since the absolute value of the tangential force is not known, the
work done upon the specimen cannot be calculated, although the
general principle of the relation of work to the plasticity deter-
minations in the plasticimeter still holds.

The method at present employed in the calculation of a plastic-
ity or working quality figure for limes or clays is as follows:

The average relative tangential force for the period of time
from zero tc five minutes is called the "plasticity figure" of the
clay or lime being tested. This method gives the most satisfactory
values in a comparison of the plasticities of materials. If the
limit of the machine is reached before the end of the five-minute
period, the remaining values are obtained by extrapolation.

Method of Testing.

Limes. — Pastes were prepared from each hydrated lime. Each
paste was covered with a damp cloth and, after standing, at room
temperature for from 16 to 18 hours, was molded into test pieces
3.15" in diameter and 1.5" high, the base being a plaster of paris
block. The test pieces and plaster of paris block were then

174

JOURNAL OF THE

placed in the instrument and revolved. As the water was ab

sorbed by the plaster, the viscosity of the paste increased and the
forces became greater — at a rate depending upon the rate <>\ ab
sorption. The relative tangential force was noted at intervals
of 15 seconds — the instant at which the operator began to mold
the specimen being taken as zero time.

In Fig. 1 are graphically shown the results secured on eleven
hydrated lime pastes, the time being plotted against the relative

1.0

09

06

01

0.6

05

04

01

02

®

"*l

®

If

©

®

/©

(9)

1 1

1 r

©

1

GT>

1 1

k

t

1
1
1

1 I

/ /
/ A

1 ,

1

1

1 I

1 /
1 L
1 It

I I

1
I

1 If
\ J

1
1

J

f )
1
l
1

I

1 / /
/ /

/-/
-7 /

/ 1

1
1
1

rY

/-~^.

^

/ /

1 //,

ftjff

_-~^

— —

s

X0.1

0 1 2 3 4- 5 e

Time in Minutes
Fig. 1 — Relation between time and relative force for hydrated limes in
the Bureau of Standards Plastici meter

AMERICAN CERAMIC SOCIETY. 175

tangential forces. Since there is no accurate method of de-
termining the degree of consistency of the pastes, three different
insistencies were used and the one having the smallest value was
:hosen for comparative purposes.

Specimens Nos. 1 to 9, inclusive, are representative of plasters
:apable of being used commercially as finishing coats on walls.
Up to a period of from 3 to 4 minutes the forces for these speci-
mens are low after which they increase rapidly. Specimens Nos.
to to 25 are not representative of commercial finishing coats,
rhe forces for these latter specimens increase rapidly from the
start, the form of the curves distinguishing them from those
>f the finishing limes. The average relative force for the period
:rom zero to five minutes was calculated for each lime and the
ower this average force, or "plasticity figure," the greater the
jlasticitv of the lime. The results of the tests on the 25 hydrated
imes are presented in Talle 1. There seems to be no doubt
hat the plasticimeter has dra .. n a distinguishing line between the
inishing and the non-finishing varieties of hydrated limes and has
ietermined their relative behavior when spread upon absorbent
;urfaces.

Lime-Gypsum. — Owing to the fact that, in the preparation of
inishing coats, burned gypsum is always added to hydrated lime,
t was thought desirable to determine the working properties of
ime -gypsum mixtures and the effect of the amount of water used
n the preparation of the pastes. An Ohio finishing hydrated
ime was mixed with amounts of water varying from 34 per cent.
:o 56 per cent., by weight. Each paste was allowed to stand
rom 16 to 18 hours and the "plasticity figure" was then de-
:ermined in the plasticimeter. To a portion of each paste was
hen added 25 per cent, of burned gypsum (by weight of the dry
lydrated lime), 0.25 per cent, of organic retarder (by weight
)f the dry burned gypsum), and sufficient water to give all of the
)astes a like degree of consistency. This consistency was made
;o approach, as nearly as possible, that preferred by the plasterer
n the coating of walls.

These mixtures were also tested in the plasticimeter. Some
)f the pastes were too soft to be molded and were placed in the
nachine in a container with the plaster of paris block at the

176 JOURNAL OF THE

bottom. In this series of tests the test specimens were one inch

thick instead of the i1 ■< inches as ordinarily used. The results

of these tests are given in Table 2 and shown graphically in Pig. 2.
The percentage water in the lime-gypsum mixtures is not taken
into account in the diagram, as these mixtures were all of pra<
tically the same consistency and water content.

Table i . — "Plasticity Figurbs" or Hydrated Limbs.

Plasticity Per cenl
No. Source, (a) Kind. figure, water used

1 Pennsylvania 'H M o. 163 42

2 Ohio '-'H M F 0.227 4.3 6

3 Ohio II M P o. 241 42 .9

4 Ohio H M F o . 244 43 . 6

5 Ohio HMF 0.248 43.4

6 Ohio HMF 0.277 44

7 Ohio HMF o . 297 42

8 Vermont :iH C 0.324 45

9 Ohio H M P o . 332 42 . 9

10 Pennsylvania H M 0.603 44-4

11 Maine H C o . 620 44 . 9

12 West Virginia H M 0.659 47-4

13 Pennsylvania H C o 663 50

14 Alabama H C 0.751 46

15 Tennessee H C 0.956 48

16 Missouri H C °-956 44

17 Pennsylvania HC 1 .01 41

18 Michigan H M 1.23 42 . 9

19 Ohio H M 1.25 46 . o

20 Tennessee H C 1.30 47-4

21 West Virginia H C 1 35 47 4

22 Virginia H C 137 50

23 Ohio HM 1.43 43

24 Indiana H C 1 . 93 48

25 Tennessee H C 326 47

'H M F — Ohio high magnesium finishing hydrated lime.
-H M ■ — High magnesium hydrated lime.
3H C — High calcium hydrated lime.

(a) Samples received and tested in 191 7.

(6) Per cent, of water to weight of the paste.

It will be noted that the curve for the lime pastes (Fig. 2) ex-
tends downward, from 38 per cent, water to the minimum plastic-

AMERICAN CERAMIC SOCIETY.

i77

:v value at 48 per cent., and then rises. In other words, the
8 per cent, mixture is the most plastic of the lime pastes. The
lost plastic lime-gypsum paste however, is that made from a
me paste containing 38 per cent, water or 10 per cent, less than
he amount required by the best lime paste. The question arises
s to why the plasterer does not use 38 per cent, water instead
f 48 per cent, in the preparation of his lime. His practice is
robably governed by the fact that a 38 per cent, mixture would
e so stiff and unworkable that the increased plasticity gained
y using more water would not offset the increased amount of
ime required in the mixing.

ABLE 2.

-Effect of Variations in the Water Content of the Pastes
on the Plasticity and Density.

Per cent.
of

water.

"Plasticity
Lime.

figures" of pastes.
Lime-gypsum.

Density
Lime.

of pastes.
Lime-gypsum

34

0560

1.56

1-52

36

O.705

O.271

1 .64

1

53

38

O.780

0.2IO

1-57

1

53

40

O.709

O.252

1-53

I

49

42

O.567

0355

1-52

1

49

44

O.527

O.304

1.50

1

49

46

0-454

0.525

i-47

1

5i

48

0.447

O.485

i-45

1

48

50

0.486

O.520

1 .42

1

48

52

0.527

0.538

1 .40

I

48

54

0.675

0.590

1.38

I

46

56

0.695

O.695

1-35

I

45

The densities of the pastes are shown graphically in the lower
art of Fig. 2 ("density," as here used, is the weight of a unit
oh.ime of the paste in grams per cc). It will be noted that the
ensities of the lime pastes decrease uniformly from a value of
.64 to 1 .35 while the densities of the lime-gypsum pastes de-
rease from a value of 1.52 to 1.45. With an increase of water
ontent the density decreases within the range of the amounts
sed — the highest plasticity not corresponding to the greatest
ensity.

Clays. — The clays tested were ground to pass a 100-mesh sieve,
uffieient water being added to produce the consistency necessary

[78

JOURNAL OF THE

for jiggering. The clay specimens tested were 3.15" in
diameter and 0.25" thick. The reduced thickness of the

specimen necessitated the use of a flat disk in place of the conical
disk. After molding, and previous to testing in the plasticimeli r,

03

0.7

06

05

04

1*0.3

01

\rPk
\ 0

istici
f Limt

ty Fiq
' Past

Lire
es

/
/
/

\

\ /

■s

■?t

/ /
/ /
J /
/

\
\
\
\

/

/

\
\
\

\
\

/
/

s

Lpi<

1

7ST/CI

Lime

ty Fie,
-Gyps

iure
jm Pa

stes

\

1

N
N

s

/
/

/

*"*»^.

'™ ' 1

v, ».//,«, v^r~

----.

Density of Lime Pastes —

1 1 1 1

I.&5

1.55

>

=—1.45

34 36 38 40 42 44 46 48 50 52 54 56> "

Percentage of Water in Lime Paste Before Adding Gypsum

Fig. 2 — Effect of consistency of lime paste on plasticity and density of
the paste alone and with 25 per cent, gypsum.

the pieces were placed in a moist atmosphere for from 16 to 18
hours. The data secured for the clays is given in Table 3 and is
shown graphically in Fig. 3.

It will be noted that the curves for the Tennessee and Ken-
tucky ball clays, which we consider as being very plastic, are
similar to those of the plastic limes while the curves of the non-

AMERICAN CERAMIC SOCIETY.

179

)lastic kaolins are similar to those of the non-plastic limes. The
elative behavior of the clays in jiggering appears to be indicated
>y the character of the curves. The action of the jiggering tool
n pressing the clay against the plaster mold is quite similar to
hat of the plasticimeter disk in pressing the clay against the
>laster of Paris block — hence the results should be comparable
n either case. Sufficient investigative work has not been done to
warrant the claim that the "plasticity figures" represent the
elative plasticities of the clays, although they appear to give a
omparative indication of the jiggering behaviors of the different
lays.

Table 3. — "Plasticity Figures" of Clays and Porcelain Bodies.

F3. F10

Clay

Tennessee ball clay

Fo.

O TQ7

Kentucky ball clay

O
O
O
O
O
O
O
1
1

56T

Illinois "kaolin"

567

Maryland kaolin

6OO

Florida kaolin

Georgia kaolin

New Jersey fire clay

682
790
87O

North Carolina kaolin

English china clay

992
12

Delaware kaolin

Porcelain bodies:

Ki (not aged)

K2 (not aged)

K3 (not aged)

K4 (not aged)

K5 (not aged)

No. 26 (not aged)

No. 26 (aged 8 months) . . .
No. 34 (not aged)

No. 34 (stored 8 months) . .
F5 = 0 to 5 minutes
F3 = 0 to 3 minutes.
F10 = 0 to 10 minutes.

1.05

0.785
0.645

0.622

0.483

857

623

937
650

Porcelain Bodies. — A number of typical porcelain bodies were
repared in the usual manner and tested in the plasticimeter.
iome of the bodies were tested at once while others were aged for

I No

JOURNAL OF THE

8 mouths previous to the testing. In the preparation of the
bodies, sufficient water to afford the desired consistency for jigger-

/ 2

Time m Minutes
Fig. 3. — Relation between time and relative force for clays in the Bureau of
Standards plasticimeter (using flat disk and quarter-inch specimens).

ing was added. The compositions of the bodies were as fol-
lows:

Body No. Kl : Per cent.

North Carolina kaolin 150

Georgia kaolin 150

Delaware kaolin 15 .0

Feldspar 20 . o

Flint 35.0

100. o

AMERICAN CERAMIC SOCIETY. 181

Body K2 — Same as Ki with 4 per cent. Tenn. and Ky. ball clays re-
placing 4 per cent, of the kaolin.

Body K3— Same as Ki with 8 per cent. Tenn. and Ky. ball clays re-
placing 8 per cent, of the kaolin.

Body K4 — Same as Ki with 12 per cent. Tenn. and Ky. ball clays re-
placing 12 per cent, of the kaolin.

Body K5 — Same as Ki with 16 per cent. Tenn. and Ky. ball clays re-
placing 16 per cent, of the kaolin.

Body No. 26. Per cent.

N. Carolina kaolin 12.5

Florida kaolin 12.5

Delaware kaolin 12.5

Georgia kaolin 12.5

Feldspar 20.0

Flint 30 . o

0.5 1.0

Time in Minutes

. 4. — Relation between time and relative force for porcelain bodies in the
eau of Standards plasticimeter (using flat disk and quarter-inch specimens).

182

JOURNAL OF THE

Body No. 34.

N. Carolina kaolin 13

Florida kaolin 13

I Delaware kaolin 13

1 ieorgia kaolin 13

Feldspar 18

Whiting I

Flint 2 s

Pei ''in
•75
■75
•75
•75
•5
•5
.0

£0.3

)l

1 ^

^ A^/

r

s

sL

i

/ 1

/ 1
/ 1
' /

II

//
//

//
//
(/

/

/

0 1

2

3

, A

l i

> i

6

i

i 10

Time in Minutes

Fig. 5. — Relation between time and relative force for porcelain
bodies in the Bureau of Standards plasticimeter (using flat
disk and quarter-inch specimens).

The results of the tests upon the porcelain bodies are given
Table 3 and are shown graphically in Figs. 4 and 5. Test pie<
1/U" thick were prepared from bodies Ki to K5, and the "plast
ity figure" represents the average relative force for the peri

AMERICAN CERAMIC SOCIETY. 183

:om o to 3 minutes. An increase of the ball clay content ap-
arently causes a uniform decrease in the "plasticity figure"
ith the exception of those for bodies K3 and K4 which are quite
lose together.

Specimens V2" in thickness were prepared from bodies 26
rid 34 and the "plasticity figure" calculated for a period of
me from o to 10 minutes. The results are given in Table 3
ad shown graphically in Fig. 5. The ageing of the bodies for

months has materially improved their jiggering qualities.

Conclusion.

It should be noted that the determinations on the clays are not
umerically comparable with those for the limes since different
eights were used on the pendulums in the testing of the two
taterials.

It may be said that the plasticimeter has given us some practical
iformation in reference to the properties of hydrated limes and
r clay bodies. There is room for improvement, however, in the
echanical construction before the instrument may be employed
correlating plasticity with the other properties of the ma-
rials.

It remains to be proven whether the property of retaining water
id the resistance to deformation under pressure is a measure

the plasticity of a ceramic body when in the plastic condition.

COMMUNICATED DISCUSSION.

W. E. Emley: I want to thank the authors for the above
iticism. If the instrument had not been criticized it would

all probability have died a natural death as so many plasticim-
ers have done before. It has now received a new lease on
e, and I hope that further experiments will bring to light the
ifects in the present design and help to perfect it.
It is unfortunate that the authors have found it necessary to
tablish certain empirical conditions for the use of the instru-
ent — such as the consistency of the paste, the thickness of the
ecimen, the duration of the experiment, etc. This makes it
fficult, or rather impossible, to compare the results obtained

1 84 AMERICAN CERAMIC SOCIETY.

under different conditions. The fact that the "plasticity figure"
has no absolute meaning blocks any attempt to correlate the
plasticity, as measured, with any of the other physical properties
of the materials. This subject is worthy of thorough and careful
investigation. Even though it should be found impossible to
design a machine which would measure plasticity in c. g. s. units,
the attempt to do so would undoubtedly contribute largely to
our present meagre knowledge of this subject.

It is very gratifying to note the definite statement that the
instrument is able to differentiate between finishing limes and
other limes. This is the purpose for which the instrument was
designed. At present the only way of making such a distinction
is by actual trial — by having the lime spread on a wall by a plas-
terer. This method is not satisfactory. It is too laborious and
the results depend upon the personal opinion of the. laborer —
who is frequently not any too intelligent and is apt to be biased.
Such results cannot be expressed in black and white for future
reference. It is therefore of great benefit to the industry to have
an instrument which can measure and express different degrees
of plasticity — even though the results are only comparative,
and not absolute.

It is very interesting to know that the instrument gives an in-
dication of the behavior of clays in jiggering. The plasticimeter
was not designed for this purpose. This would indicate that
the instrument may be adaptable to a wider field of usefulness
than has as yet been contemplated.

I glean from the present paper that the instrument is, as now
designed, capable of practical use in studying the plasticities
of limes. This should facilitate and stimulate the redesign of
the instrument to overcome the present technical faults. I hope
that the authors will continue their experiments along these lines

rHE USE OF FURNACE SLAG AS GROG IN ARCHITEC-
TURAL TERRA COTTA BODIES.

By R. H. Minton, Metuchen, N. J.

One of the most interesting and important ceramic products,
irchitectural terra cotta, has within recent years fallen into some
lisrepute on account of a few cases of failure in the building, one
:ommon type of failure being so-called "dunting." This is un-
ortunate because terra cotta lends itself to the modern system of
teel construction better than any other type of building ma-
erial.

Very little has been published on this subject. R. L. Clare1
tates that "the first and most important cause of failure is a
lefective body mixture." He also states that "experiments
vhich we have made indicate that it is much easier to produce
ire cracks or cooling cracks by a slight change in the rate of
iring and cooling than it is by changing the body composition
>ver a wide range."

In an investigation of the "dunting" of architectural terra
:otta, E. C. Hill2 states that, "fire cracking seems to depend al-
nost entirely on two factors, namely, the composition of the body
,nd the rate and method of cooling. Regarding the composition,
t was noted that bodies composed of very sandy clays are much
nore sensitive to fire cracking than those made of less sandy,
nore plastic clays." His final conclusion was that "while cooling
s of the greatest importance, the composition and density of the
>ody are large factors in the sensitiveness of the body to fire
•racking."

Hill is undoubtedly correct in his belief that the body should
>e controlled by the size and the amount of the grog, rather
han by the use of sandy clays, but he seems to have overlooked

1 Trans. Am. Ceram. Soc, 19, 593.

2 Not published.

t86 JOURNAL OF THE

the question of the kind of grog. The kind of grog is undoubtedly
of equal importance t<> the size and the amount.
Some years ago our attention was drawn to the possibility of

using furnace slag as grog in terra eotta bodies, the slag being
a material of low cost, light in weight, and possessing sufficient
resistance to destruction by the elements. Preliminary tests
which we conducted with this material were not satisfactory,
on account of the inferior quality of the slag selected, but recently
our experiments with slag for this purpose have been quite promis-
ing.

In our first experiments we used ordinary crushed slag but it
was found that granulated slag, on account of its lightness in
weight and freedom from iron, etc., offered greater possibilities.
It was soon determined that we were limited to the use of acid
slags, the basic slags being unsatisfactory as grog on account of
their excessive weight and their content of free lime which would
probably cause disintegration of the burned terra cotta bodies.
The following two bodies are typical of those prepared in our
investigation :

Body No. 174.

Stoneware clay 15 parts (by vol.)

Retort clay 15 parts (by vol.)

Slag 20 parts (by vol.)

BaC03 'A per cent.

Body No. 175.
Stoneware clay 15 parts (by vol.)
Retort clay 15 parts (by vol.)

Grog (sagger) 20 parts (by vol.)
BaC03 7s per cent.

A granulated slag from Coatesville, Pa. having the following
composition was used in Body No. 174:

Per cent.

Si02 44- 1

AI2O3 24.1

Fe203 13

CaO 20.6

MgO 6.7

MnO 3 • o

99-8

>V\ 111£

AMERICAN CERAMIC SOCIETY. 187

The bodies were prepared in accordance with the usual terra
:otta practice. Shrinkage bars, 14" X 3" X 3", and also sanitary
ray legs for glazing tests were molded from the plastic clay.
fhe tray legs were slipped and glazed with a terra cotta slip and
flaze maturing between cones 4 and 5.

The test pieces were burned in a coal- fired test kiln with cone
I turning at the end of a firing period of 45 hours. At cone 5,
3ody No. 174 develops a dark buff color and Body No. 175 a
ight buff color. Both bodies develop a good dense structure
dien fired to this cone.

The absorptions of Nos. 174 and 175 (after immersing for 48
lours in water) were 9.4 per cent, and 12.4 per cent., respectively.
"he total drying and burning shrinkage of the two bodies, Nos.
74 and 175, was 14.3 and 14.7 per cent., respectively. At the
xpiration of one year, no signs of shivering, crazing or other
'efects were noted in the white matt glaze applied to the test
deces.

Freezing Tests.

In order to best determine the relative abilities of the two
•odies to withstand weathering action, some of the shrinkage
•ars from bodies Nos. 174 and 175 were subjected to a so-called
xtificial freezing test as follows:

After boiling for three hours in a sodium sulphate solution,
he test pieces were removed and dried at room temperature for
.8 hours — the immersion and boiling treatments being repeated a
lumber of times. No apparent effect of this treatment was ob-
erved on either body up to the fifteenth test, after which Body
■Jo. 175 began to flake slightly. Body No. 174 was not affected
fter being subjected 34 times to the treatment. The condition
f the two bodies at various stages of the treatment was noted as
ollows :

Lfter 20th treatment — 13 175, flaking on all surfaces B 174 not affected.

Liter 24th treatment — B 175, entire surfaces on four sides

scaling off B 174 not affected.

iter 29th treatment — B 175, scaling quite rapidly B 174 not affected.

ifter 34th treatment — B 175, scaling off in large, thin

flakes B 174 not affected.

[88

JOURNAL OF THE

The condition, after the 34th treatment, of two of the shrinkaa
burs was as shown in Fig. 1. It will be noted that Body No.
174 was practically unaffected by the sodium sulphate test while
Body No. 175 is partly disintegrated and has the appearand
rock-faced building brick.

Fig. 1. — Bodies Nos. 174 and 175 after 34 sodium sulphate treatments.

In order to determine the resistance of the two bodies to actual
freezing, pieces of each were immersed in water until saturated,
thoroughly frozen at 50 F., and then thawed. Both bodies were
apparently unaffected after having been frozen and thawed
twenty-five times. The compressive strength of the two bodies,
before and after the natural freezing, was as follows:

Compressive strength, Compressive strength,
pounds per sq. in. pounds per sq. in.

(Before freezing.) (After freezing.)

Body No. 174 6710 6410

Body No. 175 5900 3660

It will be noted that the freezing treatments have affected bu'
slightly the compressive strength of Body No. 174, whereas the com
pressive strength of Body No. 175 was reduced about 40 per cent

Experimental Bodies.
In order to compare the slag type of body with others, a numbe
of mixtures, in which were used sagger grog, slag, slag and sagge

AMERICAN CERAMIC SOCIETY. 189

jrog mixtures, and vitrified porcelain grog as the non-plastic
ngredients, were prepared as follows:

Body No. 200 — commercial terra cotta body

No. 201 — commercial terra cotta

No. 202 — commercial terra cotta

No. 203 — Stoneware clay 15 parts (by vol.)
Retort clay 15 parts (by vol.)

Grog (sagger) 20 parts (by vol.)
BaCOs 1/j per cent, by weight

No. 204 — Stoneware clay 15 parts (by vol.)
Retort clay 15 parts (by vol.)

Grog (sagger J 10 parts (by vol.)
Granulated slag io parts (by vol.)
BaCOs V2 per cent, by weight

No. 205 — Same as Body No. 204 but containing
20 parts of slag

No. 206- — -Same as No. 204, but containing 20
parts of electrical porcelain grog

No. 207 — Stoneware clay 15 parts (by vol.)
Retort clay 15 parts (by vol.)

Grog (sagger) 16 parts (by vol.)
Granulated slag 4 parts by (vol.)
BaC03 V2 Per cent, by weight

The granulated slag used in Bodies Nos. 204 and 205 had the
Allowing composition :

Per cent.

SiOj 32.4

AI2O3 12.2

CaO 39-3

MgO 1 1 . 6

MnO 0.8

S 1.8

98.1

The slag used in Body No. 207 was from the Illinois Steel Co.,
:he composition being as follows:

Per cent.

Si02 . . . 34-2

A1203 12.4

CaO 47.5

MgO 3-6

Fe203MnO 0.9

S 1.7

100 3

190

JOURNAL OF THE

Fig. 2. — No. 202, Commercial terra eotta body.

Fig. 3. — No. 203, Experimental body, similar in composition to
commercial body.

AMERICAN CERAMIC SOCIETY.

191

Fig. 4. — Body No. 205, containing 20 per cent, slag as grog.

Fi<; ,=;. — Body No. 206, containing 20 per cent, crushed porcelain as grog.

IO.

JOURNAL OF THE

In iMgs. 2, 3, 4 and 5 is shown the structure afforded by the
different types of grog in Bodies Nos. 202, 203, 205 and 206,
respectively. The photographs were taken of a section of each
of the bodies after grinding to a smooth surface on a carborundum
grinder.

The shrinkage, absorption, and crushing strength of each body
was found to be as follows:

Per cent, burning. Per cent.

Shrinkage. (Basis absorption.

Body dry length.) Burned to Cone 4-5.
No.

IO.9
III
II. 9
1 1 O

7-2

10. 1
7.6

134

Crushing

strength.

Burned to Cone 4-5.

200

3

5

201

2

9

202

4

9

203

3

9

204

4

9

205

5

0

206

5

9

207

3

5

4600 lbs. per sq. in.

3250 lbs. per sq. in.
4325 lbs. per sq. in.
3490 lbs. per sq. in.
5300 lbs. per sq. in.
3325 lbs. per sq. in.

The results of the sodium sulphate tests made on these bodies
may be summarized as follows:

After 50 treatments.

No. 200 — Scaling considerably at one end.
201 — Scaling somewhat.
202- — Scaling slightly at one end.
203 — Flaking considerably on edges.
204 — Very slightly affected on bottom surface and edges.
205- — Only slightly affected.
206 — Flaking on the edges.
207 — Flaking considerably at one end.

The treatments will be continued until complete disintegration
has taken place. After fifty treatments bodies Nos. 205 and 206
were the least affected. It was to be expected that a body con-
taining electrical porcelain, stoneware, or a similar vitrified grog
would be more resistant than one containing fire-clay grog.
However, the vitrified grog is not obtainable in large quantities
is heavier, and detracts from the working properties of a body.

Conclusion.

If further experiments confirm the practicability of the use of
slag as grog in architectural terra cotta, an abundant material

AMERICAN CERAMIC SOCIETY. 193

t less than one-half the cost of sagger grog, and also considerably
ighter in weight, is available. In the selection of a slag, careful
ttention must be given to the source and quality — we have
ound the granulated slag to be the best for this purpose. In the
bove experiments we have experienced little difficulty from the
iresence of iron in the slag but this is a matter which would re-
hire some attention. Any iron present may possibly be removed
»y passing the slag over a magnetic pulley.

This opportunity is taken to express our appreciation for the
ssistance of Mr. E. Ogden in making the freezing and crushing
trength tests.

COMMUNICATED DISCUSSIONS.

P. H. Swalm: To the many theories for the causes of "dunt-
lg" Mr. Minton's investigation adds one to be well considered.
s the nature of the grog a factor in "dunting?" Unfortunately,
3r lack of better tests, his conclusions are based upon the dis-
ltegration or flaking of the body and his crushing strength data.
s it not a fact that the stronger bodies which resist disintegration
re also subject to the fault of sudden rupture or "dunting?"
'his has been my experience with fire clay sanitary ware bodies.

It is possible to understand why a body gives way under rapid
ooling in the kiln, and it is to be expected that a dense body,
nproperly bonded, might "dunt" under strain at some later
ime, but it seems to me that the "dunting" of a well bonded
trong body at a delayed period may only be caused, either by a
efect in molding or by the unusual physical properties of one of
he ingredients of the body. By the use of tight burning, sandy,
eavy fire clays, I have produced fire clay sanitary bodies of ap-
arently good quality, but found that these bodies "dunt" very
eadily, regardless of the size, quantity, and vitrification behavior
f the grog used. This was probably due to the introduction of
n excessive amount of silica as like results were secured by the
ubstitution of coarse sand for the fire clay grog.

The properties of the grog must therefore be considered. Mr.
linton's investigation shows that by the use of slag a stronger
nd more resistant body is produced but the effect on "dunting"

[94 l< HJRNAL OF THE

is not conclusively proven. I have found that the use of electrics

porcelain grog improved the body, but I COUld not entirely i 1 i 1 1 1

inate the sagger grog from the mixture on account of th< n
duction in strength in the raw clay state. The porcelain groj
being vitrified, is not conducive to a firm bond between the grog
and the clay. The latter point is one to be well considered in the
use of the slag as grog in fire clay bodies.

H. W. Moore: The terra cotta companies have, for at least
two years back, been trying to improve the quality of their bodiel
and the use of slag as suggested by Mr: Minton is, so far as I
know, something entirely new. I have tried a sample of slag
myself, but found it unsuitable for this purpose. If, by the use
of the slag from the Midvale Steel Company, it is possible to
produce a body which will successfully resist the sodium sulphate
tests as applied by Mr. Minton, obviously, a great stride in the
improvement of terra cotta bodies has been accomplished, as,
in my opinion, such a body would withstand any atmospheric
exposure to which it would be subjected.

E. C. Hill: Mr. Minton, in advocating the use of furnace
slag as grog in architectural terra cotta, has introduced a materia]
that is new to the industry, but which may prove of considerable
importance in the future. If slag is available at one-half the cost
of broken saggers, as Mr. Minton states, and can be obtained
free from impurities and constant in composition, its use is
certainly worth considering.

Mr. Minton refers to failures of terra cotta due to "dunting"
and suggests as a remedy the use of slag as a source of grog in-
stead of saggers. The results show that the slag body has a
much lower absorption and greater crushing strength than the
sagger body and successfully withstands the artificial freezing test
with the sodium sulphate solution while the sagger body fails
in this test. His conclusions, based on this test, are that the slag
and vitrified porcelain bodies are more able to withstand the
action of the weather than the sagger body, although he attributes
the failure of terra cotta in practice to be chiefly due to "dunt
ing-"

AMERICAN CERAMIC SOCIETY. 195

The failure of terra cotta from "dunting" is due to its inability
withstand changes of temperature, involving the expansion
d contraction of the piece, and not to its inability to withstand
(integration or flaking off, due to the destructive action of
pezing and thawing. When terra cotta fails from "dunting"
I piece is broken or cracked in sharp lines, the fracture being
gboth and cleancut; when it fails from disintegration, due to
■zing and thawing, the failure is evidenced by crumbling of the
ges and the flaking off of the sides.

The character of the failure produced by the artificial freezing
st, crumbling and flaking of the piece, would be expected from
e continued action of natural freezing and thawing but I have
ver observed the sharp hair-line cracks, so characteristic of
unting," in any of the test pieces subjected to the sulphate
>atment. I do not think that "dunting" cracks could be pro-
ced by this treatment unless the pieces were quite large, in
hch case a failure would be more likely to result from the con-
med rapid heating and cooling than from the disruptive action
the crystallized salt from the sulphate treatment.

Bureau of Standards Report.

In this connection, I would quote the following from a report
the Bureau of Standards, dated May 10, 1918, to the National
:rra Cotta Society, concerning an investigation of terra cotta
dies: "The value of this test (artificial freezing) as a criterion
the weather-resisting properties of a body is open to some
lestion. The appearance of the disintegrated specimens closely
sembles that shown by terra cotta which has failed in service,
le to alternate freezing and thawing, although similarly pre-
.red specimens have been frozen and thawed 37 times during
e past winter and do not show at all the same behavior. Only
cubes (4" solid) show any effect of the natural freezing and it
nsists of a clean, sharp break of the cube into two fragments.
9ne of the open boxes (8" X 8" X 6" deep with 1" walls) show
ects of any kind caused by the same number of freezings.
:cording to Staley,1 one sodium sulphate treatment of clay
ain tile is equivalent to two natural freezings, if the tile in the
1 Trans. Am. Ceram. Soc, 18, 642-923 (1916).

i96 JOURNAL OF THE

latter case are partly immersed in the water while freezing. Ir
the present investigation no such ratio has been obtained foj
some of the specimens disintegrated at the end of two sodiual
sulphate treatments but were intact at the end of 37 natura
freezings, a ratio of j to iS. The terra cot la specimens were
however, not immersed during the freezing but were saturated
with cold water."

Theory of "Dunting."

(E. C. Hill.)

Terra cotta bodies giving the best results in the artificial freez-
ing tests are those that are hard and dense and having com-
paratively low absorption and high compressive strength. A
body of this kind would, in all probability, be more sensitive tc
"dunting" than one less dense and would have a correspondingly
higher absorption and lower strength. In a paper read by the writei
before the New Jersey Clay Workers' Association, June, ioiy.onfire
cracking or "dunting" in terra cotta, from which Mr. Minton has
quoted, the causes of "dunting" were discussed as follows: "The
failure due to "dunting" is evidenced by sharp hair-line cracks
in the piece either occurring in the kiln or appearing later when
set in the building. The cracks are more liable to occur on large
pieces of heavy cross-section than on the smaller, lighter pieces.
When the cracks appear in the building they are more likely
to be found on pieces which are exposed on all sides than on
those set in the wall. The cause of the failure by "dunting"
is attributed to strains set up in the body during cooling. These
strains produce the hairline cracks at the time, or cause cracking
later when the pieces are subjected to changes in temperature
Experiments made show that fire-cracks may be produced in al!
bodies by rapid cooling and that, if cooled slowly enough, al
bodies are free from the fire-cracks.

Regarding the composition of the body, it appears that th<
hard, dense bodies are more sensitive to fire-cracks as are alst
those containing very sandy clays or highly siliceous fire bricl

grog-

With reference to the effect of silica in the clay and grog
data submitted by Purdy1 on the expansion and contraction 0
1 Trans. Am. Ceram. Soc, 15, 499-522 (1913).

AMERICAN CERAMIC SOCIETY. 197

orcelain body mixtures, shows that the expansion and con-
•action of fired test pieces, heated to and cooled from 900° C,

the greater the higher the flint content, and that the rate of
cpansion and contraction between 650 ° and 500 ° C is much
reater than during other stages and that the higher the flint
mtent, the greater is the rate of expansion and contraction
uring this period. The limits of composition as employed in
urdy's investigation with bodies containing clay and flint was
i per cent, clay, 80 per cent, flint and 90 per cent, clay and 10
jr cent, flint. Purdy noted that the total expansion at 900 ° C
r the high flint body is about 2V2 times that of the high clay
xly and that the rate of expansion and contraction between
)0° and 500 ° C of the high flint body is about 7 times as
•eat as the high clay body, and for temperatures below 500 ° C
)out 3 times as great. Data obtained in the same way by
oeck,1 on fired test pieces of fire clay and ball clay, shows that
le expansion and contraction curve for fire clay has the same
:neral trend as the porcelain bodies containing flint in Purdy's
vestigation, but the ball clay has a comparatively uniform rate

expansion and contraction between 900 ° and 2000 C.

If, in cooling a high clay body, the contraction takes place
; a uniform rate over a considerable temperature range,
le body becomes more sensitive to strains as its rigidity
creases, so that strains are more likely to occur at a
ill red heat and after, than at the higher temperatures. When
body contains sandy clays, however, it not only has to contract
ore in cooling, but has to withstand the greatest rate of con-
action during the period within which it is most sensitive to
>oling strains.

Since it appears that the very dense, hard bodies are
ore sensitive to "dunting" than the more open porous ones and
rice it is common practice in manufacturing many clay wares to
mtrol the density, either wholly or in part by the introduction

sandy clays, it is suggested that the density of terra cotta
)dies be controlled as far as possible by the size and amount

the grog, rather than by the use of sandy clays. It is further
iggested that the cooling be conducted slowly below 650 ° C.
1 Trans. Am. Ceram. Soc, 14, 470-479 (191 2).

i.,s JOURNAL OF THE

While the method and rate of cooling is of the greatest im-
portance in preventing "dunting," and terra cotta bodies of al-
most any density or composition will be free from dunting if
cooled slowly enough, it is possible that, with the ordinary rati
of cooling, in attempting to compound a body with low absorption,
high crushing strength, and which will stand up best in the
artificial freezing test, we might arrive at one that would be quite
sensitive to "dunting" — particularly when manufactured into
large pieces which are exposed on all sides to the action of the
weather.

In order to manufacture terra cotta with straight and true
lines it is essential that the body have as low a drying and burning
shrinkage as possible, and that there be no tendency to warp in
firing. These requirements are met with a body having a some-
what open, porous structure, rather than one which is hard and
dense.

The body for terra cotta should have a reasonably low ab-
sorption, high crushing strength and stand up reasonably well
in the artificial freezing test, but it is not advisable to go beyond
this on account of the increased shrinkage, tendency to warp in
firing, and sensitiveness to "dunting." In order to secure these
desired properties I do not believe that it is necessary to develop
a body which is able to withstand 30 of the sodium sulphate
tests.

R. H. Minton: Mr. Swalm admits my contention that the
"character" of the grog and clays used is a more important factor
in producing "dunting" than any other. This particularly ap-
plies to sandy grog or clays. His objection to the use of vitrified
grog — owing to the difficulty of bonding with the plastic clays,
is not sustained, as many terra cotta manufacturers have used
vitrified grog entirely for certain purposes, and other manufac-
turers use it in large quantities. I have not experienced this
difficulty in using vitrified grog, and even so, it would not arise
in the use of the slag which does not have the smooth, clean sur-
faces of the porcelain and stoneware grogs.

It is perhaps unfortunate that my conclusions were based upon
the disintegration and crushing strength data, but it is not clear

AMERICAN CERAMIC SOCIETY. 199

trhat other tests could have been used — as those employed are
□tost conclusive in the determination of the weather-resisting
[ualities of any clay product.

Mr. Hill states that "the failure of terra cotta from 'dunting'
3 due to the inability to withstand changes of temperature — in-
olving the expansion and contraction of the piece — and not to
ts inability to withstand disintegration or flaking due to the de-
tructive action of freezing and thawing." Mr. Hill has not
•roven that "dunting," which develops in the wall, is not pro-
iuced by the destructive action of freezing and thawing, regard-
sss of what may be the primary cause. However, it was not
itended to attempt to produce actual "dunting" by the freezing
ests. These tests and the crushing tests were applied because
hey appeared to be best adapted to a measurement of the re-
istance of clay bodies to weathering action.

Mr. Hill further states that "terra cotta bodies giving the best
esults in the artificial freezing tests are those that are hard and
lense and having a comparatively low absorption and high com-
•ressive strength ; a body of this kind would, in all probability,
>e more sensitive to 'dunting' than one less dense and having a
orrespondingly higher absorption and lower strength." Mr.
lill seems to overlook the real point at issue. The use of slag,
»r vitrified grog, does not necessarily produce a denser body so
ar as the clay matrix is concerned. The absorption is lowered
merely because the slag grog has a lower absorption than the
•orous sagger grog. One of the terra cotta manufacturers has for
nany years used vitrified chemical stoneware or electrical porcelain
;rog in the manufacture of terra cotta which he supplies to a
Canadian customer. Another manufacturer considered the use
if common stoneware for the same purpose.

Mr. Hill also states: "it appears that the hard, dense bodies
re more sensitive to fire-cracks than the others as are also those
ontaining very sandy clays or highly silicious fire brick grog."
Lnd again: "since the very dense, hard bodies seem more sensi-
ive to 'dunting' than the more open porous ones, and since it
3 a common practice in the manufacture of many clay wares to
ontrol the density, either wholly or in part by the introduction
f sandy clays, it is suggested that the density of terra cotta

AMERICAN CERAMIC SOCIETY.

bodies be controlk 1 as far as possible by a variation in the size
of the grog, rather than by the use of sandy clays."

In his original paper on "dnnting" he states: "it has been noted
that the bodies containing sandy clays are more sensitive tc
fire cracking than those made up of the more plastic clays." Also
"fire cracking seems to depend almost entirely on two factors,
namely, the composition of the body and the rate and method ol
cooling. Regarding the composition, it was noted that bodies
composed of very sandy clays are much more sensitive to fire
cracking than those made of less sandy, more plastic clays."
These statements seem to be contradictory.

My contention is that the "dun ting" and failure of the terra
cotta is dependent more upon the "character" or "quality" oi
the grog and clays than upon the other factors and that a grog
of low absorption is superior to the porous sagger grog. So fai
as price, weight, availability, working properties, etc., are con-
cerned, it would appear that the slag is to be preferred to the
grogs prepared from porcelain, stoneware, etc.

THE EFFECT OF ELECTROLYTES ON SOME PROPER-
TIES OF CLAYS.1

By H. G. Schurecht, Columbus, Ohio.

Introduction.

One of the sources of loss in the manufacture of clay products,
ipecially of the larger kind, chemical stoneware, glass pots, etc.,
due to the lack of strength of the clays in the dry state and at
ie burning temperatures. Sufficient strength to resist the shocks
id strains of handling in the green state is necessary. At cer-
tin periods in the burning the ware loses strength and is liable
> warp, crack, or break, causing large kiln losses. The natural
:medy would be to improve the strength of the clay body by
'tificial means or to introduce stronger clays.
The effect of the addition of organic compounds to clays has
len studied by Acheson,2 Minton3, Seger and Cramer,4 and
!oerner,5 and it has been shown that the strength of the clays
in be increased by the addition of gallo-tannic acid, catechu,
raw emulsion, dextrine, or starch.

The effect of the addition of alkalies and acids on the volume
irinkages of clays in the plastic state has been reported by
leininger and Fulton,6 and Kerr and Fulton,7 but our informa-
on as to the effect of these electrolytes on the strength of clays
meagre. Bleininger8 has reported that cast bodies containing
kalies may be stronger than bodies not containing alkalies. The
riter9 has noted that the rate of disintegration in water is much
ower for the deflocculated than for the flocculated clays.

1 By permission of the Director, Bureau of Mines.

2 Trans. Am. Ceram. Soc, 6, 31-64 (1904).

3 Ibid., 6, 231-259 (1904).

4 Wein und Leipzig, A. Hartleberi s Verlag, 1909, p. 82.

5 Op. cit., pp. 82-84.

6 Trans. Am. Ceram. Soc, 14, 827-839 (1912).
' Ibid., 15, 184-192(1913).

8 Bureau of Standards, Tech. Paper 51, 40 (191 5).

9 Trans. Am. Ceram. Soc, 19, 144 (1917).

JOURNAL OF THE

A-> the rate of its disintegration is sometimes used as a rougl
indication of the bonding power of a clay, further investigation
with a view to studying the effect of the addition of alkalies ant

acids on the strengths of clays were carried out. A surprising
increase in strength was gained by the addition of the alkalies
in sonic cases the strength of the treated clays being four time!
as great as that of the untreated. The present investigation wa*
limited to a study of the effect of the addition of alkalies and aeidi
on the properties of clays with special reference to their strength

Method of Testing. — For the investigation the following mix
tures were selected:

(i) 50 parts Kentucky ball clay and 50 parts flint.

(2) 50 parts Tenn. ball clay and 50 parts flint.

(3) 50 parts Georgia kaolin and 50 parts flint.

(4) 50 parts Kentucky ball clay and 50 parts grog.

(5) 100 parts Kentucky ball clay.

The cross-breaking strength, water of plasticity, drying shrink-
age, density and porosity were determined for each clay in the
raw condition. The tests were made on the untreated clays
and on the clays to which 0.05, o. 10, o. 15, 0.20, 0.40, 0.60, 0.80,
1 . 00, 1 . 50, 2 . 00, 3 . 00 and 5 . 00 per cent, (by weight) of electro-
lyte had been added. The electrolytes added were sodium
hydroxide, sodium silicate, sodium carbonate, calcium hydroxide,
tannic acid, and sulphuric acid.

Tests of the Clays in the Raw Condition. — For the cross-
breaking tests, bars 1 " x 1 " x 7 " were molded and dried at room
temperature for 3 days, then in a drying oven at 65 ° C for 24
hours, and finally in the oven at no° C for 24 hours. A span
of 5 inches was used in breaking the pieces transversely. The
modulus of rupture of each piece was calculated — the average of
ten determinations being taken as the final result.

In making the water of plasticity and the drying shrinkage de-
terminations, the electrolyte was added to a quantity of water
slightly less than that required for developing the maximum
plasticity. In case the limits for the normal consistency were
large, the water was added by volume so as to keep the per cent.

AMERICAN CERAMIC SOCIETY. 203

Ided as nearly as possible the same, thereby making the shrink -

je data more comparable.

The drying shrinkage was determined by the use of the kero-

ne immersion method and the per cent, volume shrinkages were

ilculated on the basis of the dry volumes.

The per cent, shrinkage in terms of the true clay volume was

itermined by the method described by Bleininger1 as follows:

d(v\ — Vo)
s = — — — 100.

w

s = per cent, volume shrinkage in terms of the true clay volume.

d = specific gravity of the powdered clay.

vx = volume of the clay briquette in the wet condition.

V2 = volume of the clay briquette in the dried condition.

w — weight of the dried briquette.

Tests of the Clays after Burning. — The test pieces were burned
1 cone 01 in a gas-fired test kiln, the temperature being in-
eased at the rate of 30 ° C per hour from 10000 C to within
>° C. of the finishing temperature — at which point the rate was
duced to io° C per hour. The apparent porosity, volume
Lrinkage, and density of each briquette was determined.
The cross-breaking tests were made on 4" X 0.8" X 0.8" bars,
sing a span of 3.2".

Effect of the Electrolytes on the Properties of the Clays
and Mixtures.

Dry Strength. — In Figs. 1 and 2 is shown the effect of the addi-
on of the electrolytes on the dry strength of the clays and mix-
ires studied. The addition of the alkalies and tannic acid to
le clays increased the strength while the addition of sulphuric
:id decreased the strength. Calcium hydroxide, when added in
nail quantities (0.2 to 2.0 per cent.), improves the strength
miewhat but weakens the bodies when added in larger amounts.
he maximum effect is produced by the addition of from 0.8
) 2.5 per cent, of the alkalies, or about 3.0 per cent, of tannic
;id.

1 Bureau of Standards, Tech. Paper 1, 21 (1910).

204

|< HJRNAL OF THE

When caustic soda was added the bodies were- molded with SOlfl
difficulty and would not retain their shape unless a low pei cea

NrJ

>»

/

■■o

Ken </,•*(, Bat' Clay
Fi.n >

u

3
3

J.?

/

-

O

-

v*.

iPj

-'

a
•a

-.'

- _

. ^_

___'a

■

Hi !

~~

§

-V"

'i-

~~

■- —

~-i^-

-" "

."~

...

--"

-"

—

— "

0

Per cent, electrolyte in terms of clay.
Fig. i.— Dry Strength.

2000

1

l

5

- 5

o *<4 Ball Cla
o xy Ball Clac

f Ontj SO OrOtj
1 and 50 F/in t

6

/

f

SO Go Kaol'i ana SO Fltnf-

3

/

/

"o

/

,■''

l""

-..^

1

f

/

**«

■^

I

/

/
/•

-

~^_~

.--

---

s

I

f-f

y

*■—

--

~^

"

-■"

4O0

1 1

r

I'

--^

' 1

-r-

-

■'

—

(7

1

•<

<■

3

Per cent. NaOH in terms of clay.
Fig. 2.— Dry Strength.

of water was employed, although in this condition the clays lamir
ated badly. This effect was most pronounced in the mixture

AMERICAN CERAMIC SOCIETY.

205

^•eloping the greatest strength (see Fig. 1). The addition of
lium silicate, sodium carbonate, or more than 2 per cent, of
phuric acid, affected the molding properties to a smaller ex-
it and the addition of calcium hydroxide or tannic acid had
:le effect on the molding properties. The addition of less than
3 per cent, sulphuric acid was found to improve the working
jperties of a body.

flie natural brown color of the ball clays changed to black
on the addition of a large percentage of alkalies or acids. The
; of calcium hydroxide did not materially affect the natural
oring of the clays although it did cause the formation of flaky
lies on the surfaces of the test pieces.

Under the microscope the clays treated with the alkalies ap-
ared to be very fine grained, indicating that the clay substance
»sent was in a deflocculated condition.

Water of Plasticity. — -The curves in Figs. 3 and 4 show the
set of the electrolytes on the per cent, water of plasticity. In
» case of the two ball clays the addition of the alkalies decreases

1

..nil-

--

- —

-: '

.

so

fen tuctttj
fhnt

Ball C/oif

. 40

,y

/

-

,*-

^~S

5<

S?4

■■J

— 33

c

i4'r"

->

....

--

—'

lv

\

rij

COj

~ \

--J

--

--

— ~

-----

' ~

v

**"**

" "

V

---

1 jii

)}-'■

\

cd

i ,fl

\

-"I

0

^

La

u

PL| 25

\

\

_--

--

—

-~"

to

»

Per cent, electrolyte in terms of clay.
Fie. 3. — Water of Plasticity.

206

JOURNAL OF 'I'll

the per rent, water necessary to the development of the- plastii
The addition of a small amount of acid necessitates the use of a
greater amount of water in working the clays but, if large quanti
ties are added, a smaller amount of water must be used. The
addition of calcium hydroxide increases the per cent, water of
plasticity.

1 1

SO rfy &ail C/atj and so Groo

\

o

\

.2

"3. 54

>

\

V

.

-

e
u

\--

*!.

1...,

- —

u

\
\

\

1 "'

-"

—

-

—

. _

-

X

S

__-

--

--

"~ -

'-_.

Per cent. NaOH in terms of clay.
Fig. 4. — Water of Plasticity.

Volume Shrinkage. — In Figs. 5, 6 and 7 is shown the effect of
the electrolytes on the volume shrinkages of the clays and mix-
tures. The addition of one per cent, of alkalies appears to de-
crease the volume shrinkage of the mixtures of 50 parts clay and
50 parts flint with the exception of the one containing Georgia
kaolin, the shrinkage of this mixture being increased. The addi-
tion of sodium hydroxide to the Kentucky ball clay increases the
shrinkage. Small percentages of sulphuric acid increase and
large percentages decrease the volume shrinkages. The addi-
tion of calcium hydroxide and tannic acid increases the volume
shrinkage.

The ratio of pore to shrinkage water becomes lower as the per

AMERICAN CERAMIC SOCIETY.

20:

1

55 /rentac/r^ Bo// C/au
so nih*

,'

eg,

r--

_

___-_-

.-

—

--

- -

-

_.-

■ '■- -

-^

rtu<

--

~.

/fl

' •

V\

^■*

f

y4

Ir !

1/Vl'C

_4fi.

[J

-

— -

•

7

y

'

./

'

O-y

..--'

Per cent, electrolyte in terms of clay.
Fig. 5. — Volume Shrinkage.

rzrz '

iy 5a// Clay

0 Ku #<7// (T/^y and SO

6roq

:

0 /ru 0a// Claw and So r/m't
0 Tenn Ball Claw ana' SO flint

/>■-.

-•"\

.•-"

— -

-'

h\

/'

w.

.-=».

— -

.■==.-.

-----

S=-«

^1

>

/-

_.

—

--

N

=--

—

—

--

- —

/

Per cent. NaOH in terms of clay.
Fig. 6. — Volume Shrinkage.

208

JOURNAL OF THE

cent, of shrinkage water in terms of the total water increases until
2.0 per cent, of caustic soda has been added when the revei
true. As clays having a low pore to shrinkage water ratio and a
high per cent, of shrinkage water in terms of the total water often
developed good strength, these results are in accordance with the
strength data. (See Fig. i.)

1

/

V

/

/

So Ke/Uucity Sail Clotf
SO fl-nt

,

/

1 1 1 1

\

f —

Slnririkaqe Water

V

\

\

Terms of Total Water

\

\

\

I-

'

.

i '»'

\

!

'<

\

.—

\

— '

0

1

I

t

5

Per cent. NaOH in terms of clay.
Fig. 7. — Shrinkage and Pore Water Relation.

Density. — In Figs. 8 and 9 are shown the effect of the electro-
lytes on the densities of the clays and mixtures. The alkalies
and tannic acid increase the densities of the bodies, the maximum
effect being produced by the addition of from o . 40 to 3.0 per
cent. When more than 2.0 per cent, sulphuric acid is used the
density is increased to a remarkable extent. However, there is
only a small increase in strength (see Fig. i), which may be
accounted for by the oxidizing effect of the sulphuric acid and the
consequent destruction of the organic bond. Additions of cal-
cium hydroxide up to 1 . o per cent, increases the density. The
addition of larger amounts decreases the density.

AMERICAN CERAMIC SOCIETY.

209

-50 to ^t/c*V 8a,i ctai
SO Fhnt~

tC

NoOr<__

/

~~

/

'

'

t»

/

N*&9-

i

/

-

'

r

;,

-'

- ■

■ —

7-

'

/

/'

^N

1

/

'

1

1

no

I'1

1 . »"'

i. 1 1

rrj

A.-

___

__ -

- -5

.-_

"-

—

k.

^r;

T"

us

■~--*«s^

-.

""~~

--.

•-

-

— ■

Per cent, electrolyte in terms of clay.
Fig. 8. — Dry Density.

1

8a II Clay

'

f '

"n

So ffif Ba/i c/ay 0"d so Groa

SO Tenn 8a 1/ Cloy a"d SO Win t

/

....

s

iO (

1<7 rtaolm and SO f//nf

/

"

~~

-;-

1^-

--_

/

»

"

>.

j

l'\

'

'35

U

Q

'S

-/■

'

ISO

- i

1

140

0

i

%

S

Per cent. NaOH in terms of clay.
Fig. 9. — Dry Density.

Porosity. — In Figs. 10 and n are shown the effect of the elec-
■olytes on the porosities of the clays in the dry condition. The
orosity is decreased by the addition of the alkalies while the ad-

2[()

JOURNAL OF THE

.»■

---

—

—

-

_

30 Kentucky Boll Oau
SO Hint

rd&

.-

-

v

-.<-—-\Z\

•>.,

f

_.&SS*

■&

■■-.

■„ -

-N

•

Ye

L<Oi

\

Hat

—

\

•

■

Ma On

\

I

\

\

_

—

1

H

t

i

Per cent, electrolyte in term-; of clay.
Fir-. 10. — -Dry Porosity.

Ph

1

50

„.„„

C/ay and so Flint-
Cfoy and -50 Grog

a\

^0

rew* <&7// r/oy <7/^ 50 F/mr
Oo Kaolm and SO Flint

i 'A

1 n\

l>

..-

~-

~~

ft

\

_

,-"

v;

H

r'

.'

;

v*

\

^

"J

— -

\ \

/'

s

,*

,''

\

y \

,,-''

-

—

h-

— _,

\

\

\

/

Per cent. NaOH in terms of clay.
Fig. ii. — Dry Porosity.

AMERICAN CERAMIC SOCIETY.

211

tion of the acids in small percentages increases and in large
:rcentages decreases the porosity. The porosity is increased by
e addition of sodium hydroxide.

Burning Properties. — In Fig. 12 is shown the effect of sodium

-■^

-*0
10

u

-"-

•"-■"

__£

fns.

k_

__

--

"'

SO Kentucky Bali Cloy
50 Flint

s

y

Burned to Cone Ot

s

y

•

1

L*V _

70

r

IS

10

—

(

falu

T«:

.,.,-«

life

5

fv~

--.

.__

— ■

n

f

Per cent. NaOH in terms of clay.
Fig. 12. — Burning Properties.

droxide on the burning properties of the mixture of 50 parts of
mtueky ball clay and 50 parts of flint when burned to cone 01.
le volume shrinkage decreases with the addition of small quan-
ies, up to 0.8 per cent., and is increased by the addition of
■ger amounts. The density is increased and the porosity de-
based by the use of the sodium hydroxide. Considerable
imming and surface vitrification is noted with the addition of
ge percentages of sodium hydroxide.

In Fig. 13 is shown the effect of the electrolytes on the strength
the mixtures when burned to cone 8. Although the fluxing
tion of the alkalies increases the strength, there is a decided
tiilaritv between the burned and the raw strength data (see
£. 1 J. In the burned condition, the maximum effect is pro-

I' fl RNAX OF THE

duced by the alkalies when fromo.8to 2.5 per cent. is added -

the caustic soda being the most effective.

Jlgi

„

1 _.

\

/

1

•

~

■,

j

j

- —

"

"\ '

■

--

,<t--

-.

i

\

- -

-P'\

"*

!

\

■auw.

—

—

—

~Y~

—

.-

\

ic

> Kentucky Boll Cloy

3

urned to Cone 3

Per cent, electrolyte in terms of clay.
Fig. 13. — Burned vStrength.

Summary.

The strength of clays when dry may be increased by the addi
tion of the following electrolytes to the bodies in the plastic
state, the first named being the most effective: (i) sodium hydrox-
ide, (2) sodium silicate, (3) sodium carbonate, (4) tannic acid, and
(5) calcium hydroxide. The addition of sodium hydroxide in-
creases the strength of some of the mixtures as much as 400 per
cent. Small percentages of sulphuric acid decrease the strength
of Kentucky ball clay. More than 2.0 per cent, of calcium hy-
droxide decreases the strength of Kentucky ball clay.

The addition of from o . 80 to 2.5 per cent, of caustic soda al
fects adversely the molding properties of the bodies, as evidenced
by their inability to retain their shape when molded, unless a
very low per cent, of water is used. This is also noticeable, to a
lesser extent, with additions of sodium silicate, sodium carbonate.,
or more than 2 .0 per cent, of sulphuric acid, but is not so notice-

AMERICAN CERAMIC SOCIETY. 213

>le with additions of calcium hydroxide or tannic acid. The
klition of less than 2.0 per cent, of sulphuric acid improves the
olding properties of the bodies.

In the case of the two ball clays, the addition of the alkalies
wers the per cent, water of plasticity. The addition of the acids

first increases and then decreases the water required for work-
g the bodies. The addition of calcium hydroxide increases
e water of plasticity.

The addition of from 1 .0 to 2.0 per cent, of alkalies decreases the
)lume shrinkages of the 50 parts clay and 50 part flint mixtures
ith the exception of the one in which Georgia kaolin is used, in
is case, the shrinkage being increased. The addition of sodium
•droxide to Kentucky ball clay increases its shrinkage. Sul-
mric acid, in small percentages, increases the shrinkage and in
rger percentages decreases the shrinkage. Tannic acid and
Icium hy droxide increase the shrinkage.

The addition of alkalies and acids increases the densities of
e clays and mixtures when dry. In amounts up to 1 .0 per cent.
Icium hydroxide increases the density. The addition of larger
nounts decreases the density.

The porosity is lowered by the addition of alkalies and in-
eased by the addition of acids in small percentages. Targe
Iditions of acids are accompanied by decreases in porosity. Cal-
um hydroxide increases the porosity.
The burning shrinkage at cone 01 is lowered by the addition

small percentages of caustic soda and is increased by the addi-
3ii of larger quantities. The addition of sodium hydroxide in-
eases the density and decreases the porosity in the burned con-
tion .

The addition of the following electrolytes to the bodies in the
astie condition increases the strength of pieces burned to cone

the most effective being named first: (1) sodium hydroxide, (2)
•dium silicate, (3) sodium carbonate, and (4) tannic acid.

Discussion.

Mr. Whitakkk: I would like to ask Mr. Schurecht what
feet ageing would have on the strength of the bodies to which

JOURNAL OF THE

the electrolytes have been added' Would ageing still further ii
crease the strength '

Mr. Schurbcht: We have not determined the effect oi agl
iiiK on the strength of the clays, although we have determine
the effect of the ageing of clay slips to which electrolytes have bee
added and found that it changes the viscosity. The ageing ur
doubtedly has an effect upon some of the other properties of tli
clays.

Prof. Stagey: It would appear that one of the chief advai
tages in having a clay of good strength would be that we migh
expect less cracking during drying. Several years ago, I di
some work on "The Effect of the Addition of Salts on the I )r\
ing Properties of Clays."1 In this I noted that the addition c
the alkalies invariably increased the tendency of the clays t
crack in drying. Possibly the pieces of very carefully dried wan
free from cracks, would have shown high strength, as in the cas
of Mr. Schurecht's samples, but it does not appear to be goo)
practice to increase the possible strength of a clay by the add]
tion of alkalies, if these additions are going to introduce dryin;
difficulties, which make it very difficult to obtain sound wrare ii
ordinary factory practice.

In my wrork, I have found that the introduction of small amoi
of acid and neutral salts decreased the tendency to cracking.

mat

Mr. Shaw: My experience in the use of electrolytes in th
casting of bodies bears out that of Mr. Staley. The tendenc
to crack was very great. I should add, howrever, that the plasti
clays which I used were of very low grade.

COMMUNICATED DISCUSSION.

W. W. Greenwood: Mr. Schurecht's paper presents materi:
which is of interest to the chemist as well as to the ceramis
In dealing with so complex a subject as the chemistry of electa
lvtes, great care of manipulation is essential if the results of tl
work are to be of much value. In this particular piece of uoi
the writer shows such care and has handled the necessary detai
1 Trans. Am. Ceram. Soc, 17, 697-711 (1915).

AMERICAN CERAMIC SOCIETY. 215

his experiments with a thoroughness which makes one feel
:e accepting his observations as authentic.
It would have been helpful, however, if in this paper the author
fuld have gone a little more into detail as to the theory of his
bjeet or if he had speculated as to why the phenomena which
■ observed occur. This is of value for two reasons: first, be-
.use in most cases the experimenter is better able to correla te
s work with existing information — he being more familiar with
.e subject than the outsider; second (and this applies to the case

hand), because by not doing this Mr. Schurecht leaves the reader

a position where he may draw wrong conclusions from the facts
esented.

In his summary, he states: "The dry strength may be increased
; adding the following electrolytes to the bodies in the plastic
ate, those which are most effective being named first: (1) sodium
rdroxide, (2) sodium silicate, (3) sodium carbonate, (4) tannic
:id, and (5) calcium hydroxide." There is danger in assuming
Dm this statement that the electrolytes mentioned were directly
sponsible for the increased strength of the dried bodies. There

nothing in the observations presented to uphold this theory.
1 fact, it would seem quite probable that the electrolytes are
lly indirectly responsible for this action. We know that their
Idition de flocculated the clays and the statement is made in the
iper that "under the microscope it could be seen that the clays
eated with alkalies consisted of very fine grains, a characteristic

deflocculated clays, while the untreated clay is much coarser."
oes it not seem reasonable that the increased strength observed
is at least partially due to the fineness of the particles in the
sated clays? These finer particles would tend to pack closer
tether and consequently there would be fewer voids than in
e untreated clays. In other words, this resulting strength is of
echanical origin brought about by chemical means. If such is
e case — and there is much in the paper which points in this
Section — the electrolytic action is simply a means towards an
d — a method if you please of obtaining fine particles of clay
mpacted together which in turn produce greater mechanical
rength.

2i6 JOURNAL OF THE

In another pari oi his summary the statemenl is made TM
strength after being burned to coik- 8 may be increased by addinj
the following electrolytes !<> the bodies in the plastic state, thoi
which are most effective being named In i (i) sodium hydroxiq

(2) sodium silicate, (3) sodium carbonate, and (4) tannic- acid.j

Here again there is a chance for difference of opinion as to tin
cause of this increased strength. No doubt it is of chemical origin
but there is nothing in the paper to show whether the reagenj
named produced it because they were electrolytes or becau
their alkali content. Binns,1 in discussing the action of tannii
on clays, notes that alkaline salts in the tannins would becon
operative at the higher temperatures.

How would one of these treated clays, when dry, differ from a
untreated sample? Would there not be minute particles of alk
line salts about each grain? When the clay was burned these sal
might be expected to exert a fluxing action and consequent
a stronger and denser body would result. In the familiar procc
of "salt glazing," the alkali, freed from the decomposing sodiu
chloride, acts on the outside of the ware and glazes it — here tl
alkali would be in a position to act as a fluxing agent throughout
the structure of the ware. So again the electrolytes themselvJ
would be only indirectly responsible for the results noted.

The work of this paper indicates another bit of data which i
would be worth while to determine. In two cases the order ■
effectiveness of these electrolytes is (1) sodium hydroxide, (2
sodium silicate, (3) sodium carbonate and (4) tannic acid. It woul
be of interest to know whether this is due to the characteristics 0
the reagents themselves or to their alkali content. In the tw,
cases cited, the order of strength as electrolytes is directly propoi
tional to the alkali content. Sodium carbonate would, in solutior
hydrolyze and show a weak alkaline reaction; sodium silicate ad
as a fairly strong alkali toward litmus — depending of course upo
its concentration — and sodium hydrate is the strongest of all||

It may be possible that the whole theory of the effect of electa j
lytes on clays is really based upon the action of alkalies upon clay
1 Trans. Am. Ceratn. Soc, 6, 57 (1904).

JOURNAL

OF THE

.MERICAN CERAMIC SOCIETY

A monthly journal devoted to the arts and sciences related to
? silicate industries.

.1. 1 April, 1918 No. 4

EDITORIALS.

HE PRESENT UNUSUAL DEMAND FOR MEN WITH
TECHNICAL CERAMIC TRAINING.

Probably the best known of the trite sayings of Andrew
irnegie is the admonition to "put all your eggs in one basket
d watch that basket" but another which has enjoyed a very
de circulation is the advice given to a business friend which
is, in substance, "If you have a hundred thousand dollars to
end on the development of a new proposition, spend the first
ty thousand in finding out what is already known."
Twenty years ago, or at about the time the American Ceramic
ciety was formed, it was in rare cases only that any systematic
tempt was made to find out in advance of embarking upon a
w proposition or of attempting to develop some new product or
w manufacturing process, what successes or what failures had
companied similar earlier experiences in attempting to aecom-
sh the same or similar results and it was almost never that a
stematic attempt was made to determine in advance what laws
nature, chemical or physical, were to be met with in carrying
t the undertaking and what influences such laws must have in
termining the limitations under which the work must be done.
Today the situation is very different. The evolution has been
Host a revolution. The old methods of secrecy and distrust
d ignorance have given way to frank discussion and to more
an a beginning of real knowledge in the science of silicates.
; an integral part of this evolution (or revolution) has come the

218 JOURNAL OF THE

establishment and growth of departments in our institutions d
learning devoted to the study of the fundamentals underlying aj
processes of silicate manufacturing and the dissemination through
out the industry of the knowledge thus accumulated.

This evolution has borne fruit abundantly. With it there ha
come an ever-increasing demand for men with suitable training
capable of taking their places in the manufacturing industries
This demand is felt in all branches of the silicate industries — in th
clayware industries, in glass plants, in cement mills, in enamelin]
plants and in the many allied manufacturing establishments.

With this logical development of the industry along norma
lines has come the added stimulus of war demands. One featur
of the very unusual situation created by the world war is th
necessity that has arisen in many branches of the Ceramic In
dustries, in common with almost all other manufacturing in
dustries, of replacing with materials of domestic origin or o
domestic manufacture, materials which were formerly imported
Such importations have been cut off either because importei
articles formerly came from the enemy countries or because th
great demand for shipping facilities for purely war purposes ha
made it impossible to continue shipments of materials whic!
could be duplicated at home. It has taken a tremendous amoun
of investigation and research, both theoretical and practical, t
accomplish the substitution of suitable domestic materials for th
former imported articles.

Another special demand upon the Ceramic industries is tha
for new products and improved forms and qualities of old product
made necessary by the new demands of a war business. Th
Nation's business is, and has been for a year and a half, the bus:
ness of War and products demanded by the new manufactun
incident thereto have had to be made. New knowledge and ne
experience have been most essential in meeting the new dimcultu
involved.

With a careful analysis of the situation at hand, it is not difficu
to understand the present great demand for men trained in tl
sciences upon which the silicate industries are founded. Oi
institutions of learning report very heavy demands for men, ar
these demands they are quite unable to satisfy with the cor

AMERICAN CERAMIC SOCIETY. 219

)arativelv small number of men undergoing such training. Erom
he larger public and private laboratories interested in the silicate
vork come the same reports of numerous demands for trained
nen. The demand far exceeds the supply. It cannot be denied
hat this situation of heavy demand and limited supply has been
esponsible for the fact that training has not always been as
•omplete as it should have been, but very great strides have been
nade, within the past decade especially.

The calling is one offering a very attractive future in the in-
eresting nature of the work, in the possibilities of development,
ind in remuneration and there can be no doubt that in proportion
o the increase in a general realization of this situation will come
tn increase in the number of young men who fit themselves
o supply this demand by entering our colleges and universities
o gain the required training in the scientific courses dealing with
•eramic work.

AMERICAN CERAMIC SOCIETY AT THE FOURTH

NATIONAL EXPOSITION OF CHEMICAL

INDUSTRIES.

The American Ceramic Society will be represented at the
7ourth National Exposition of Chemical Industries to be held at
he Grand Central Palace, New York City, the week of Sept.
•;,rd.

The participation of the Society in the Exposition of last
ear evidenced an unusual interest in ceramic work and this year
he Exposition management has arranged for the grouping of the
■eramic exhibits in a Ceramic Section. The American Ceramic
Society will have its headquarters in this Section.

The Ceramic industry includes that branch of the chemical
ndustry engaged in the production and manufacture of silicates,
ncluding as it does a group of industries producing materials and
vares to the value of hundreds of millions of dollars, it is one of the
argest single branches of manufacture based upon chemical
echnology.

Although the public at large has perhaps appreciated the
ecent progress made in other branches of the chemical industrv,

AMERICAN CERAMIC SOCIETY.

such as in the manufacture of dyes and munitions, it does not navi
a thorough realization of the extent "i the ceramic industries do

of their basic importance to the so called "essential industries.'
The war essential nature of some of the products which are beinj
produced by this group of manufacturers should also not he over
looked.

Among the ceramic products which are vitally necessary to tin
successful manufacture of other essential products we vvouh
mention fire brick, furnace linings, flue linings, crucibles, retort
and other refractories used in the production of iron, steel
brass, copper and other metals and their alloys; abrasives to
the grinding of optical products, munitions, machine parts, etc.
chemical porcelain and stoneware for the production of acids
heavy chemicals, explosives, etc. Portland cement for ships
building construction, etc.; building and construction ma
terials as brick, tile, fire-proofing, terra cotta, conduits, sewd
pipe, drain tile, etc., necessary for the construction of Govern
ment buildings, ordnance storehouses, training camps, etc. ; anc
Enameled iron ware used in the manufacture of chemicals
explosive products, etc., and very extensively for the serving o;
food in both the Army and Navy.

The Ceramic Industry has also been called upon to develoj
and produce certain military products of a vitally essential nature
and which were imported into this country previous to the war
Among these should be mentioned optical glass for gun sights
range finders, periscopes, etc. ; special glass for searchlight
reflectors, ships' lights, etc.; spark plugs for airplane engines
improved refractories for the furnaces of destroyers and othei
battle craft; and laboratory glassware and porcelain fo;
experimental and control work in the chemical industries.

The Ceramic Section at the coming Fourth National Expositioi
of Chemical Industries will undoubtedly awaken the interest — no'
only of those engaged in the production of ceramic wares — but als<
of many of the engineers, chemists, superintendents and other
in attendance and in whose work imporant use is made of Cerami
products.

ORIGINAL PAPERS AND DISCUSSIONS.

:he effect of the degree of smelting on
the properties of a frit.

Hv K. P. Posts and B. A. Rice, Elyria, Ohio.

Introduction.

n a consideration of the actual composition of a frit, we as-
le that certain constituents entering into the raw mixture
s through chemical changes which bring about the volatiliza-
i of certain portions and the fusing together of the remaining
■edients into a homogeneous mass having a certain definite
mical composition. This we express in terms of the melted
les or by means of the molecular formula. Not only is this
unption made relative to the enamel frits, but it also finds
lication in connection with fritted glazes, glasses and various
tr ceramic products which result from the use of mixtures
uch materials as soda, feldspar, borax, fluorspar, etc.
fpon reading the literature on the subject of the chemistry of
L we find that the above general conditions are taken for
ited; we do not find a report of an investigation leading to a
iparison between the actual composition of the resulting frit
the theoretical composition as shown by the computed for-
a.

I we assume that, at some particular stage in the smelting of an
mel, the actual composition corresponds to the theoretical
iposition, there must be two stages, one on either side of this
ticular point, during the first of which there are still present
ie of the materials which are volatile in the ordinary sense of
term, and beyond which point further heating may possibly
se decomposition and volatilization of some of those ma-
ils which we ordinarily consider as being non-volatile.

JOURNAL OF THE

Some months ago, a series of experiments was started wli
lias thrown considerable light on tliis particular subject and

results may be of sufficient interest to others engaged in
ceramic industries to warrant the presentation of this paper.

Experimental.

The work has included the melting of a commercial enai
formula under conditions covering a large variation in the ti
of smelting. The enamel thus produced has been studied,
as to its behavior when applied to the ware, but as to the ch<
ical and physical changes which have been brought about by
different degrees of smelting. Samples of each have been ac
yzed and the results compared with the theoretical composit
of the frit. Deformation tests have been made in order to
termine the effect of the various degrees of smelting upon
deformation behavior of the resulting frit. Tests to determ
the effect of the varying heat treatments on the resistance of 1
enamels to acids have also been made.

As to the formula used, it will be sufficient to state that it \
fairly resistant to acids and prepared from the usual enar
constituents. We can draw definite conclusions in a relat
manner without stating the formula of the enamel or the act
analyses of the resulting frits.

The raw materials, properly mixed, were placed in a tilt
smelter which had been pre-heated in the usual manner,
soon as the material had reached a fairly uniform condition
fusion and after thorough stirring, a small portion of the enai
was withdrawn from the smelter — the total time required
reach this condition being, taken as unity. This enamel ^
designated as "A." In like manner samples were poured fi
the smelter at the following time intervals based on the fonj
as unity: i l/z, 2, 21 2, 3 and 31, 2, the samples being designated
B, C, D, Eand F, respectively.

The samples thus prepared were thoroughly dried and ijj
formly ground dry, without mill additions, in an experime: 1
pebble mill. That portion of the resulting powder passin i
200-mesh sieve was used as the basis of the three following te i

AMERICAN CERAMIC SOCIETY. 223

1. Chemical Analyses.

.Tie chemical analyses of the six samples were made in the
lal manner. The scheme outlined by Landrum1 proved very
pful in this work. Mellor's Treatise on the Ceramic Indus-
s, Vol. i, proved a most reliable guide.

2. Deformation Test.

L portion of the 200-mesh powder was mixed with water and
organic binder and molded into test cones of standard dimen-
is. The cones were thoroughly dried and mounted in pairs
a piece of asbestos board. This mounting was placed in an
:tric furnace (at a temperature of 600 ° F.) in such a manner
t the junction of a thermocouple was just between the two
es. The temperature of the furnace was gradually increased at
niform rate and two points noted. The first of these was the
iperature at which the tip of the cone was first observed to
ve and the second, the temperature at which the tip of the
e was just touching the asbestos board upon which the cones
v mounted. By noting both cones in this manner and tak-

the average temperature at which they were first observed
deform and at which they were touching, two readings were
ained, the first of which was termed the starting point and

second, the deformation point. The term "Deformation
ge" has been applied to the interval existing between these
> temperatures and has been considered as indicating the
eral range through which the enamel is in a semi-fused condi-
1 but still sufficiently viscous to remain in place.

3. Acid Test.

'he acid test was made on the 200-mesh powder in accord -
£ with the methods which have been previously worked out
I reported.- The method in brief is as follows:
'wo grams of the frit were carefully weighed in a 400 cc.
ker, to which was added 100 cc. of the desired
ition (10 per cent, hydrochloric acid in this case).

! Trans. Am. Cenun. Soc, 12, 144 (1910)

-Ibid, 17, 137 (1915); 18, 570 (1916); 19, >4r> (I9I7)-

JOURNAL OF THE

The contents of the beaker wire thoroughly stirred, covered will
.1 watch glass and then allowed to stand at the laboratory leni
perature for 24 hours. The remaining frit was then washed int<
a weighed Gooch crucible, all traces of acid removed, the crud
ble dried and weighed, and the loss of weight determined. Thl
loss of weight was taken as the acid loss under these conditions
In this manner, by running a series of tests under parallel condi
tions, a very good comparison results. Obviously, the actual leni
perature has an effect on the results but under these particulai
conditions the results are comparative as the same tempera
ture has existed during all of the tests.

Chemical Analyses: Results.

As to the actual analyses of the six samples — the first step
was to determine whether fluorine was present in the frit as £
considerable quantity of fluorspar had been used in the rain
batches. Thorough qualitative tests did not detect the presence
of fluorine. This fact is interesting in view of the statemenl
made by Landrum1 — "In a ground coat enamel for sheet steel,
where the frits are properly fused, the fluorine from the calcium
fluoride entirely disappears. In white enamels, however, when
cryolite is added as an opacifier, the melting is not carried sc
far and fluorine does remain in the enamel." The enamel in
question contains no cryolite and the fluorine from the fluorspar
was all volatilized — even in the first sample which had been
smelted to a lesser degree than it would be in the production oi
an ordinary' commercial frit from this formula.

An appreciation of the results of these analyses can be had
by a consideration of the Si02, AbO.-j, B203, CaO, K20, and Xa2C
contents. There were other ingredients present in small quanti
ties in the enamels but their fluctuations throughout the serie
are not important.

In Table I we have taken the theoretical composition of th«

enamel as a standard and have indicated the excess or deficiency

of Si02, AI2O3, B2O3, CaO, K20 and Na20 as related to the theo

retical amounts. For instance, in enamel "A" there was presen

1 Trans. Am. Ceram. Soc, 14, 544 (1912).

AMERICAN CERAMIC SOCIETY

225

.88 per cent, less of Si02 than is called for in the theoretical
•rmula while in "B," there was 0.46 per cent. more.

Table I.1

Si02 —3-88

AI203 +0.51

B203 — 0.10

CaO +0 73

K,0 +1.63

Na20 +1.18

B.

+0 46
+O.40
— 1 .80
+O.52

-f- 1 . 10
+0.18

C.
+ 1 -32
+O.50

2 .90

+ 0.37
+ I OO
1 . 22

I).
+ 2.56

+ C)

—3

+ 0
+ 0

E.
+4 .62

+0.62
— 4 . 20
+0.30
+0.40

F.
+ 7 92
+0.80
— 6.00

+0. 17
+0.26

—2.77

A graphical interpretation of these results is very interesting.
1 Fig. 1 we have considered the horizontal line at the center of
le chart to indicate the actual determination corresponding

Fig. i.

► the theoretical value. If the analysis value exceeds the theo-
tical amount, it has been designated by a position above the
^rizontal line, and if the figure has been below the theoretical

1 In three cases the values here reported are slightly different from the
lalytical data — modified to be consistent with the balance of the data.

22b J< iiRNAl, OF THE

amount, it lias been placed below the Line. The horizontal d
tances AB, AC, AD, A.E and A.F, indicate the time of smelti
of each of the respective samples, A being taken as unity.
A study of Fig. i leads to certain very definite conclusions. T

changes prior to "A" cover the period during which the water
vaporized and most of the nitrates, carbonates and fluorides i
decomposed. Beyond "A," a very definite series of changes
another type take place.

It will be seen that the increase in the time of heating has pi
duced a very rapid increase in the percentage of Si02 up to "B
after which the rise is very positive but more gradual.

The Al2().t content has not been subject to very marked chang
At "A," it was slightly above the theoretical amount and throughc
the series it has at first dropped and then increased gradual
and at "F" is somewhat above the value at "A."

The CaO content has slightly decreased from "A" to "I
although always being slightly above the theoretical amount.

The percentage of alkalies has dropped quite materially frc
"A" to "F" and it is particularly interesting to note that t
volatilization of Na2() has been much more rapid than that
K20. In fact, throughout the series there is more K20 than t
theoretical amount, while the Na20 content drops below t
theoretical shortly beyond "B" and at "F" there is a deficien
of nearly 3.0 per cent.

The B203 content is about equal to the theoretical at ■
and then drops off steadily to "F."

There appears to be a definite point at "B," at which the co
position of the entire batch is nearest to that of the theoreti
composition — in fact the algebraic sum of the values is practica
zero. The line "X" has been drawn through this point on
of the charts. It would seem that the most radical chang
due to the volatilization of substances ordinarily considered
volatile, have come practically to equilibrium at this point. 1:
water has been completely driven off, the nitrates decompos
the carbonates broken down and the fluorine — with its equival' I
of silica — has vaporized as silicon tetrafluoride. In addition
this there has been a noticeable loss in B203 — due to volatili
tion.

AMERICAN CERAMIC SOCIETY. 227

Beyond this line "X" the volatilization of B2()3 continues at
out the same rate although there seems to be a radical change
the type of the curve for SiC>2 and Na20. It is quite apparent
it the volatilization of the B203 and Na20 is the determining
•tor in this change and that the increase in the Si02 content
the frit is the result.

It is further very interesting to note the point corresponding
the time at which this enamel would normally be poured. The
e "Y" has been drawn through this point. We have the fol-
ding values: Si02 +17, Al2Oa +0.5, CaO +0.3, K20 +19,
l20 — 1.5, and B2O3 — 3.0. In other words, the frit, as it
uld normally leave the smelter, is considerably higher in Si02,
newhat higher in K20, A1203 and CaO, slightly lower in Na20,
:1 very materially lower in B203 than would be indicated from
! theoretical composition obtained from computation.
The data in Table 1 has been calculated on a percentage basis
i as is characteristic of data of this kind — a change in the
lount of any one constituent, when the actual amount of the
lers remains constant, makes a difference in the percentage
lues of the other constituents.

■\n effort has been made to present the analytical data in a
m which will eliminate these objections (Table 2). In doing
s we have assumed that, in view of the absence of fluorine in
," the actual amount of Si02 left in the batch through the
jsequent stages of the smelting is the same, and that the changes
ich have taken place have involved the volatilization of other
)stances.

Table 2.
a. b. c. d. E. F.

>i02 Assumed constant.

U2O3 o —0.56

hOi o — 3 . 70

-aO o — 0.57

C2O o — 1 .01

o — 2 . 30 — 3 .

\ssuming that in heating from A to B no change was made in
: actual amount of SiQ2, the values for the other oxides have
in changed, keeping their ratio to the Si02 the same and bring-

.56

—0.71

— 0.

95

—5 -70

—7-

■77

-^.89

— 1 .

■19

—1 .80

— 1 .

•73

—3 -92

-5 ■

73

-085

22

—9 -SO

05

—1-34

95

— 2.31

55

— 6.30

JOURNAL OF THE

ing the SiOa to the same actual amount ;is was indicated in "A.
In view of the fact that the percentage <>f SiQj was constantly in
creasing from A to F, the assumption that the- actual weight ha

been constant lias caused a decrease in the actual amounts c
the other oxides present. In columns B, C, I), F and I*, th
actual loss of Alj().t, BjOs, CaO, K2() and Xa2() as compared wit
the actual amounts present in A has been expressed. For in
stance, in B, assuming that there was no loss of Si02, the amoun
of B2().t is 3.7 points less than was present in A

In Fig. 2 the data of this table is presented graphically. Th
line at the top of the chart is designated as zero. The dill (ten
times of smelting are indicated by A, B, C, D, E and F. The dis

A

s

c

0

£

/■

^

*=■_

r

_^

\

>

, V

\

' 5 <•

2— <

)

<

k.

s "

■ 1

1

1

X.

\

.4

.3

V.

\

'■--

h"~<

*.

)>

'**«.

"""-<

)

■<

x

)

(

J Al2 0, -

Y

2 CoO

3 B?0,

4 Ha,0

5

f>#

Fig. 2.

tance below the zero line indicates the loss in the amount of tr
various constituents as compared with the amounts present in f
From Fig. 2 it will be evident that, on assuming a constai
amount of Si02 from A on, there has been a gradual loss in tl
amount of CaO and AI2O3, a greater loss in K20, and a very m;
terial loss in the amounts of Na20 a id B203. The change in tf

AMERICAN CERAMIC SOCIETY. 229

lirection of the various curves at the line "X" is again quite
loticeable.

The above data and curves would indicate that there is a loss
>f CaO and AI2O3 due to the continuous heating from A to F.
f this condition is taken as improbable, the data could be re-
irranged, keeping either the content of CaO or AI2O3 constant,
n which case the value for the SiO-> would constantly increase.
U the same time there would still be a definite decrease in the
imount of BjO.j and the alkalies.

As a general conclusion, from the data presented from this
loint of view, it is quite clear that there is a very marked volatiliza-
ion of the BgOa and alkalies. Further than this there has been
ither an increase in the amount of Si02 or a decrease in the
imount of CaO and AI2O3. It hardly seems possible that an ab-
orption of SiOa from the walls of the smelter could have taken
>lace and no other source is possible. We therefore conclude
hat a slight volatilization of the other two oxides took place.

Results: Deformation Tests.
The results of the deformation tests which were made in ac-
ordance with the foregoing outline have been rearranged as
hown in Table 3. The deformation temperature of A has been
aken as a standard and the increase in the deformation tempera-
ure above this point is the figure which is given as the value for
he deformation temperature of any particular enamel. The defor-
nation range, as given, is the actual temperature interval occurring
>etween the first perceptible tipping of the cone and the final touch-
ng of the tip of the cone at the base of the cone mounting.

Table 3.

A. B. C. D. E. P.

)eformation Temp.. o° F. io° F. 30 ° F. 500 F. 6o° F. 80* F.
)eformation Range... 60° 8o° iio° 1200 1400 1700

The above figures are readily interpreted by the use of Fig. 3,
vmich shows both the deformation temperatures of the enamels
i, C, D, E and F, relative to the deformation temperature of A,
md their actual deformation ranges.

A study of Fig. 3 throws some light on the effects of the chem-
cal changes, as indicated by the analyses, upon the physical

JOURNAL OF THE

properties of the enamels as indicated by the deformation points
and deformation ranges. It will be noted that the curve for the
deformation temperatures approximates a straight line which
rises quite uniformly from A to I\ the total increase in deforma

S. '»

<5> 60

Z Oeformahon Temperatures

1

Zs

s

'

/

/

1/

s

s

s'

y

\

r

s-

/ ^

1

Fig. 3.

tion temperature being about 80 ° F. The curve for the deforma-
tion range is of a very similar nature, showing a range of 60 ° F.
at A and about 1700 F. at F.

The observations on both of these curves having been made to
the nearest io°, allowance of this amount for experimental error
would make either of these curves very uniform.

There are two additional physical properties of the frit which
it will be interesting to take into account. By drawing a hair
of enamel from a batch during smelting there is a period during
which the hair will contain particles of unfused material and will
be quite brittle. As the mass reaches a quiet fusion, the brittle-
ness of this hair greatly diminishes and finally a point of maxi-
mum toughness is reached, beyond which it again becomes brit-
tle. For this particular enamel this point of maximum tough-
ness would be crossed by the line "Y" in Fig. 1.

A further observation has to do with the coefficient of expansion
of this particular enamel. The theoretical enamel has a coefficient

AMERICAN CERAMIC SOCIETY.

231

linear expansion of 93.7 X io-7, calculated by the use of the
ctors given by Hovestadt.' The coefficients of expansion for
, B, C, D, E and F have been computed from the analytical
ita. These data are given in Table 4 and are plotted graph-
ally in Fig. 4.

Table 4.
Coefficient of Linear Expansion.

Theoretical v B. C. D. E. *.

93.7 X io~T 101.1 96.9 92.3 88.4 88.0 88.2

v>

105

V

Thevef'ca/

V

93.7 « W 7

>

.

)

Y

to

ABC D £ F

Fig 4. — Coefficients of expansion.

Results: Acid Resistance Tests.
The tests for acid resistance under standard conditions gave
:sults as shown in Table 5.

Table 5.
A. B. C. l). B. F.

Acid loss 0.97 g. 0.65 0.57 0.38 0.24 0.14

These figures express the actual acid loss and are the inverse
f acid resistance. They are plotted in Fig. 5. It will be noted
1 Hovestadt (Jena Glass), page 217.

23 2

J< >URNAL OF THE

that. although in "A" nearly i .o gram of the pulverized frit
was dissolved, in F slightly loss than o. i.s g. was dissolved, indi
eating a very marked gain in acid resistance

e \-

5 _S .

4

;:

3

V

Fig. 5. — Acid loss

Conclusions.

In general, we may conclude that the chemical composition!
of a frit at the point of normal pouring may differ quite ma-
terially from the composition which would be expected based on
theoretical computation. The principal differences in the case
at hand are that the Si02 content is materially higher, the AI2O3,
CaO and K20 are slightly higher, and the Na20 and B203 con-
tents are very much lower than the theoretical composition would
indicate. Prior to this point the batch as a whole has passed
through a stage approximating the computed composition with
the exception of the B203 content. Further heating produces very
marked chemical changes which rapidly increase the Si02 con*
tent and decrease in like manner the content of Na20 and thei
B2O3, when judged from a percentage point of view. Or, as-
suming the Si02 to remain constant after the volatilization of
the silicon tetrafiuoride, there is a slight volatilization of CaO,
AI2O3 and K20 and very marked losses of B203 and Na20.

As a result of these chemical changes, we introduce an objec-
tionable rise in the maturing point of the enamel but a favorable
increase in the deformation range. The computed coefficient of

AMERICAN' CERAMIC SOCIETY. 233

linear expansion drops quite rapidly and at the end of the series
it is no doubt sufficiently low to produce shivering. The enamel
would have a marked tendency to over-burn and become very
brittle when applied to the ware, were the smelting to be continued
very far beyond the line "Y." The further smelting would pro-
duce a very desirable increase in the acid resistance of the enamel,
but it would appear that the resulting objectionable physical
properties would make it impossible to realize this increase in
chemical resistance.

The Klykia Enamei.rd Products Co.,
Ei.yria, Ohio.

COMMUNICATED DISCUSSION.

J. B. Shaw: Although the results obtained by the authors
of tliis paper could be more simply expressed by giving the actual
formula used, this method of presenting the data is logical and, if
carefully considered, reveals practically the same information as
could be secured from a study of the results in connection with the
formula itself.

Although I do not believe that the use of this method of pre-
senting data should be encouraged — where possible authors should
give complete data if their papers are to be of the highest value —
I do believe that in case the formulas cannot be given, some
method such as the one employed should be devised. Data
presented in this manner does not lend itself readily to criticism.
However, there are a few points of interest which should be
emphasized.

1. The points of the highest acid resistance and the greatest
toughness are not coincident in the case of the enamel used. In
other words, if great toughness is desired, it is necessary to
terminate the fritting just before the point of the highest acid
resistance is reached. On the other hand, it is necessary to con-
tinue the fritting beyond the point of greatest toughness if high
acid resistance is desired. My experience with a number of
enamel formulas has confirmed this.

2. The remarkable decrease in B203 accompanying the increase
in Si02 appears unusual. To my mind, it is doubtful whether the
E$2< h volatilizes to the extent indicated under the short heat

234 JOURNAL OF THE

treatment. 1 have found that, in making the chemical anal] -

of enamels, the B.!< >.■-, and Si()_, arc- separated with difficulty and,

unless jgreal care is exercised, some of the- ]'>■•()■.■ may be included
with and reported as Si( )...

It would add very materially to our knowledge if a reliable
method for the chemical analysis of enamels containing fluorides,
boric acid, SbjO.,, Sn< >■., etc., were available.

3. I believe that the influence of the coefficient of expansion
as a factor in shivering or crazing is greatly overestimated by
enamelers. Fish-scaling is the bete noir of the enameling in-
dustry— especially of the sheet iron enameling branch. This
phenomenon is presumably parallel to that of the shivering of
glazes. It is evidenced by the scaling off of small flakes of the
enamel and a resultant pitted surface — the steel being exposed
in many places. It is supposed to be influenced largely by the
coefficient of expansion of the enamel but in a recent set of experi-
ments— involving the use of 120 ground coats of widely different
chemical composition and applied to light steel — not one piece
showed any fish-scaling when the enamel coating was properly
applied and fired. But when improperly ground, applied or
burned, or when applied on very heavy steel, fish-scaling was very
prevalent. These facts, together with similar experiences in the
past, lead me to believe that the mechanical manipulation of the
enamel is a far greater factor in controlling fish-scaling than is the
chemical composition.

R. R. Danielson: The authors of this paper present some
interesting facts to the enameler — in that they very forcibl}
bring out the effects of undersmelting and oversmelting upon th<
enamel batch. Their conclusion that "variations in the prop
erties of the enamel may be due to the volatilization of certaii
constituents in the batch" is instructive and may account for th
irregularities which occur in the working of enamels — in spit
of all the precautions taken in the processes of preparation an
application.

Some of the irregularities in the data of this paper may bt
due to the method of sampling employed. It is quite possibl
that, in selecting a sample of undersmelted enamel, there migl

AMERICAN CERAMIC SOCIETY. 235

E certain unfused refractory constituents, such as flint, etc.,
hieh would not pulverize as quickly as some of the other sub-
:ances in the ball mill and which might fuse together to form a
)ft glass. In screening the sample through a 200-mesh sieve —
nless the grinding had been very thorough — the coarser ma-
:rials might be left on the sieve and consequently be eliminated
om the frit. This would be especially true of the undersmelted
lamels — since there would not be the homogeneity necessary
1 properly smelted enamels unless the entire sample were ground
ery finely and then quartered for the tests. This objection
ould also apply to the tests for acid resistance as carried out by
le authors.

The authors have started a work which should be continued,
>r it will undoubtedly help to solve some of the problems
icountered in enameling.

H. F. Staley: The losses by volatilization, over and above
lose ordinarily assumed, are certainly high compared to those
icountered in smelting enamels for cast iron. At the point
V," at which the authors state this enamel would normally be
oured, the loss, assuming the Si02 content to be constant from

on, is 1 1.8 per cent. From a series of carefully kept records
{tending over a period of years, I found that the loss in the
nelting of cast iron enamels exceeded the theoretical loss by
.9 to 3.0 per cent, and that part of this loss could be attributed
) mechanical loss in the filling and emptying of the smelter,
he difference must be largely due to the difference in composition
{ the two types of enamels. In the ordinary enamel for cast
on, used in this country, the B203 content varies from 6 to 12
er cent. The enamel used by the authors evidently contained a
trger quantity of B203 for at "F" (Table 2) the loss by volatiliza-
on was 9.5 per cent., a considerable portion remaining in the
itted enamel.

It is noteworthy that the B2():i and the bases were volatilized

1 approximately equal molecular amounts. This is in accord

ith results obtained by myself when working with borate glasses.1

1 these glasses, when the molecular ratio of base to B2O3 equaled

1 "The Viscosity of Molten Glasses," Eighth Intern. Congr. Appl. Chetn.

236 JOURNAL OF THE

or exceeded i : 4, the ratio of the most acid pyrogenetic borati ,

there was no loss of BjOg without loss of the base ill the ratio
existing in the melt. When the ratio of base to h>(), was lower
than 1 : 4, there was a corresponding loss of B2O3.

\i. P. POSTS: Mr. Shaw's first point is very well taken and as a
general conclusion has been confirmed a number of times. He is
slightly in error, however, in his reference to the maximum acid
resistance of this enamel. It is shown in Fig. 4 that the acid
resistance at F is relatively high but that it has apparently not
reached a maximum. In other words, had the smelting been con-
tinued a further increase in acid resistance would have resulted.

( )ur experience in the analysis of the boro-silieates does not
confirm that of Mr. Shaw. Our B20:i determinations were made on
separate samples, the silica being determined in the usual manner.
Fluorine was not found present. We were aware of the possible
errors arising in the determination of B2O3 and A1203 and ob-
served the necessary precautions as noted in Mellor's "Treatise on
Ceramic Industries," Vol. 1, pages 584 and 589. Had any ap-
preciable error of this nature occurred — -the analyses would have
totaled materially over 100 per cent. This was not found to be
the case.

The remarks relative to the effect of the coefficient of expansion
as a factor in fish-scaling are very interesting. Our recent ob-
servations have caused us to lose faith in the coefficient of ex-
pansion theory for fish-scaling — at least when the coefficient of
expansion is alone considered and computed in accordance with
the factors given in the Jena researches. We have come to the
conclusion that, although this may be one of the factors, there are
certainly other more important ones and principally those of
elasticity and tenacity as has been so clearly pointed out by
Staley.1

The criticism of Mr. Danielson is very pertinent — in so far as it
applies to the method in general. It would perhaps be preferabl
to grind the entire sample so that it would pass a given mesh
sieve rather than to take only that portion which had passed
through the sieve. It is doubtful, however, whether the errors
1 Trans. Am. Ceram. Soc, 14, 523-531 (1912).

;

AMERICAN CERAMIC SOCIETY. 237

resulting from this source could have materially affected the
results in this case for it has been our observation that the amount
of material retained on a 200-mesh sieve under the particular
conditions of grinding was very small. The point is well worth
keeping in mind, however.

The figures given by Mr. Staley — relative to the loss on smelt-
ing and in excess of the theoretical amount — would indicate a
material difference in the types of enamels under consideration.
Certain commercial formulas, in use at the present time, show a
loss of from 3.0 to 5.0 per cent, in excess of the theoretical amount
which we had formerly accounted for in the same way that Mr.
Staley accounts for his losses of from 1.9 to 3.0 per cent. We
conclude that some of this loss was due to volatilization. The
enamel used in our experimental work was considerably softer
than the ones concerning which commercial smelting data is
available. No doubt the difference in the constitution of the
enamels accounts for the apparent differences in the amount of
material volatilized.

A COBALT-URANIUM GREEN GLAZE FOR TERRA COTTA.

Bv IIkwitt WILSON, Columbus, Ohio.

We were confronted with the problem of developing a commer
ciaJ green glaze having a deeper shade than any of the yellow-
greens we had produced by the use of chromium stains in glazes
fired to cones 6 or 7.

The characteristics and methods of producing green colors in
glazes may be outlined as follows :

1. Copper greens are used extensively. At cones 6 or 7, how-
ever, the greater part of the copper is volatilized and the re-
sultant shades of green are light and uneven.

2. Chromium and chromium stains when used in most glazes
of the Bristol type give greens which appear yellow when com-
pared with deep greens. The chromium can only be used in glazes
in which the zinc content is very low. A reduction ot the zinc
content necessitates either a higher maturing temperature for
the glaze or the use of lead compounds or frits. Chromium com-
pounds are unstable in the kiln. In the proximity of even small
amounts of chromium, glazes containing zinc oxide fade to brown
and glazes containing tin are apt to be streaked with pink.

3. The nickel green colors are variable.

4. Iron compounds, if used alone, are volatilized at cones 6
or 7.

5. Uranium compounds, if used alone, produce variable yellow-
green colors.

6. Greens may be produced by the blending of blue and yellow
colors.

(a)' It is reported that mixtures of iron and cobalt compounds
have produced good greens. This has not been our experience —
our trials giving browns and brown-greens.

(6) Orton notes the preparation, by roofing-tile manufacturers,
of a gr^en sulphate from antimony and cobalt oxides. Various
yellow stains — some of which derived their yellow color from

AMERICAN CERAMIC SOCIETY. 239

antimony — were blended with cobalt and produced yellow and
greenish browns which shaded into blues.

(c) Mixtures of cobalt oxide and titanium compounds are
reported to have produced greens.

(d) Several investigators1 have produced greens from mix-
tures of soluble uranium and cobalt compounds which were added
to the glaze as insoluble frits.

In our attempt to develop the shade of green desired the pro-
cedure was as follows:

Series A.

Using a dull matt glaze as a base, uranium oxide and a cobalt
stain were added to this glaze in varying amounts.

The formula of the glaze was as follows:

o 261 K..O ]

0.306 CaO I

0.294 ZnO f 0.384 AI2O3 2.08 Si O2

0.083 BaO J

0057 MgO J

This glaze matures at from cones 3 to 7 in a commercial kiln.
The blue stain added to the glaze had the following composi-
tion:

Per cent.
(Raw).

Cobalt oxide 16 o

English ball clay 53 o

Zinc oxide 31.0

The commercial uranium oxide was supplied by the Carnotite
Reduction Co. of Chicago. A rectangular series of 45 members
was prepared by the addition of the cobalt stain and uranium
oxide to the matt glaze. The corner glazes of the series were
as follows:

Glaze No. 1 — Matt glaze + 7.0% uranium oxide + 0.0% blue stain
Glaze No. 9 — Matt glaze + 70% uranium oxide + 6.0% blue stain
Glaze No. 37 — Matt glaze + 3.0% uranium oxide + 0.0% blue stain
Glaze No. 45 — Matt glaze + 3.0% uranium oxide + 6.0% blue stain
1 F. H. Riddle, Trans. Am. Ceram. Soc, 8, 210 (1906); R. H. Minton,
Ibid., 9, 777 (1907); B. S. RadclifTe, Ibid., 16, 209 (1914); La CSramique, 1907,
No. 225, S. 25; Tonind. Ztg., 1907, S. 141 2.

•1"

|( >URNAL I »!•' THE

Each glaze was sprayed upon 3" X 5" terra cotta tile to which
a coating of white underslip had been lirst applied.

The glazed 1 iKs were placid in a commercial terra cotta muffle
kiln and burned to between cones 6 and 7. The burning period
was 70 hours and the cooling period 80 hours.

The appearance of the glazes of this series is noted as follows:

(Mazes Nos. 1, 10, [9, 28, and 37, containing no cobalt stain,
developed the characteristic uranium colors. 3.0 per cent, of

Uranioirn Oxide Per

cent.

7.0 6.0 SO

4.0 3

0

1

10

19

28

3!

2

II

20

29

3e

3

12

21

30

33

Vl ?.?5

k

4

13

22

31

40

% 3.0

^ 37S

5

14

23

32

4/

3

4*

6

13

24

33

42

7

lb

25

34

43

e

17

26

3S

44

la

IB Z7 36 4$

Series A Maff Glaze as Base
Series B Briqht Glaze as Base

the stain produced a brownish yellow color, changing to green
as the uranium content increased. The glaze containing 7.0
per cent, uranium oxide, although green, had a decided yellow
cast.

Glazes Nos. 2, n, 20, 29 and 38 were of deeper green shades
and showed the influence of the stain but slightly.

AMERICAN CERAMIC SOCIETY. 241

Glazes Nos. 3, 12 and 21 were as green as any of the glazes of
the series, but were not uniform when applied on large surfaces.

Glazes Nos. 30 and 39 were mottled by patches of yellow.

Glazes Nos. 4 to 9, 13 to 18, and 22 and 23 were uniformly
deep green in color.

Glazes Nos. 24 to 27 developed a distinct blue-green color.
The green color faded and the blue became stronger with in-
creasing amounts of the cobalt stain.

Glaze No. 40 contained the smallest amount of the cobalt
stain to show blue tints. When the uranium addition was low-
est (3.0 per cent.), the blue shades were obtained with the low-
est amounts of blue stain.

The glazes listed above as good deep greens were somewhat
defective on account of the occurrence of yellow patches on the
surfaces. There was no regularity in the occurrence of this yel-
low surface mottling. This mottling was very prominent in
several of the good green glazes which were applied to full sized
pieces of terra cotta. The appearance of the mottling was not
very objectionable — in fact was pleasing — although not desired.

Series B.

Using a bright glaze of the following formula as a base, another
series was prepared by the addition of the uranium oxide and cobalt
stain in the same proportions as in Series A:

Bright (".laze:

0.238 KjO

0.354 CaO

0.210 ZnO \ 0.428 AI2O3 2.90 S1O2

o. 144 BaO

0.055 MgO

This glaze matures at from cones 4 to 7 when fired either in a
small or large kiln.

The appearance of the glazes of Series B is noted as follows:
The colors developed were in general the same as those of
the Series A glazes, although not as distinct. The green colors
tend toward the brown and drab shades. The blue and yellow
colors are less pronounced. The addition of uranium oxide alone
tends to produce fine dark specks.

JOURNAL OF THE

The yellow mottling developed by some of the glazes of Series
A was not observed in any of the glazes of Scries 15. The mottling
of the glazes of Scries a may be explained as follows:

Since the uranium compound is partially soluble, a greater
proportion of the yellow material was deposited near the surface
of the glaze as the water evaporated in drying. The yellow color
of the uranium partly masked the blue color of the insoluble
stain and produced the mottled effect. Although the uranium
does not eome to the surface as a scum, it tends to take possession
of the surface and the upper portion of the glaze — -leaving a
greater proportion of the blue stain next to the underslip. Any
particles of blue stain which do appear on the surface are coated
with sufficient uranium to appear green. Under the microscope,
a cross section of the glaze coating shows that in places where
the yellow-green does extend into the mass of the glaze a darker
and bluer green results close to the underslip. As the bright
glaze used in Series B produced no mottling, we may assume
that it had greater solvent power than the matt glaze used in
vSeries A.

Series C.

From the information secured from Series A and B a third
series was prepared as follows:

In order to avoid the mottling of the matt glaze, a soluble blue
stain was employed — the cobalt being added as soluble cobalt
sulphate. Maintaining the uranium content constant at 6.0
per cent., the cobalt sulphate was added to the matt glaze (Series
A) in varying amounts. The addition of 3.0 per cent, gave a
reliable color, the proportion of uranium to cobalt sulphate in
this case being 2:1. The matt glazes to which the cobalt sul-
phate was added were more vitreous than those in which the in-
soluble cobalt stain was used.

No signs of mottling were in evidence in the glazes of series C
and the 3 . o per cent, cobalt sulphate glaze was used in the coat-
ing of 200 sq. ft. of terra cotta surface. A green glaze of this de-
scription can be burned in proximity to other glazes without
danger of discoloring them.

By calcining or fritting the cobalt and uranium to a homo-

AMERICAN CERAMIC SOCIETY. 243

geneous mixture and by subsequent fine grinding a similar green
can no doubt be obtained.

COMMUNICATED DISCUSSION.

R. L. Clare: i. I wish to state that whenever possible we
have avoided the use of uranium compounds in terra eotta glazes
on account of their excessive cost and we have usually been
able to produce the various shades of green desired without the
use of these materials.

2. We have used chromium alone in glazes and have found it
to be very unstable and unsatisfactory. It, however, in some
cases gives a very deep green color, deeper and better in fact
than any which we have been able to secure by the use of a stain.
It appears that the materials used in a stain with the chromium
tend to dilute the green shade. Materials such as iron, cobalt
or uranium when used in a stain change the shade to a brown-
green, blue-green or a yellow-green.

3. It is very interesting to note that Mr. Wilson has produced
a good green by the use of a mixture of uranium oxide and a cobalt
stain. It is not clear in my mind whether his resultant green
is of the same shade as the typical chromium green. We have
produced greens by the use of mixtures of iron and cobalt stains
and manganese and cobalt stains, but in each case they have de-
veloped yellow-green, blue-green or brown-green colors and in
no case were the typical deep grass-greens of the chromium pro-
duced.

Mr. Wilson's theory of the cause of the mottling on the glaze
surface is very interesting and plausible. We have often noticed
scumming on the surfaces of some of our colored glazes and we
have attributed it to the solubility of some of the colorants and
the consequent concentration of salts on the surfaces of the
pieces — due to evaporation. It has already been shown that colors
can be produced by painting or spraying certain soluble salts in
solution on the surfaces of the unburned terra eotta glazes.

D. F. Albery: I wish to note the following in reference to
the methods of producing green colors in terra eotta glazes:

1. Copper oxide is satisfactory in glazes fired from approxi-
mately cone 04 to cone 2.

244 J< >URNAL OF THE

2. Chromium alone is too unstable to be recommended for the

production of the green colors.

3. Chromium stains when properly compounded, calcined and
washed, produce very stable and uniform greens of practically
any desired shade. These green shades may be varied from deep
blue-greens to brown greens by the addition of cobalt oxide, iron
oxide or burnt umber. The composition of the stain influcm <
its color and shade considerably. For instance, it is possible to
vary the shade of a given glaze by substituting plaster of paris
for whiting in the stain. The composition of the glaze also has
a decided influence on the stability of the color produced by a
given stain.

We have made name panels by using a green glaze for the back-
ground and a white glaze — containing about 6 per cent, tin oxide —
for the raised letters. The white glaze did not show the slightest
discoloration even though it was in contact with the green glaze
in a line at the base of the letters. It is largely a matter of glaze
composition. Our green glazes are not fritted and do not con-
tain zinc or lead.

4. I have never known of an iron compound producing a green
color when used alone in a glaze at any temperature.

5. I have produced an unstable yellow-green by the introduc-
tion of a high percentage of uranium.

6. I have produced good blue-greens with cobalt-rutile com-
binations in glazes maturing at from cone 03 to cone 2.

Although good green colors may be obtained by cobalt-uranium
combinations, it would appear that the cost of the high per-
centage of uranium oxide necessary would discourage the exten-
sive use of this method.

R. H. Minton: Those who have had experience in the manu-
facture of architectural terra cotta will appreciate the value of a
dependable green coloring agent which is not affected by zinc
oxide, and which will not affect, or be affected by, other colors.
White, greens, browns and tans are widely used in polychrome
work and a combination of these colors has always presented
difficulties if the green was produced by the use of chromium
compounds — the usual practice. Terra cotta white matt glazes

AMERICAN CERAMIC SOCIETY. 245

usually contain a high percentage of zinc oxide and, if used next
to a chromium green, an unsightly line of tan or brown results
at the juncture. For this reason, a green produced from a mix-
ture of uranium and cobalt greatly simplifies the factory opera-
tion.

My experience with the use of uranium compounds for yellows
and greens has covered its use in both fritted and raw, bright
and matt glazes, and has shown that in order to produce a good
yellow from uranium in a matt glaze, the zinc oxide should be
.'liminated as it tends to the development of greens. I have
found that in order to produce uniform uranium colors, either
yellow, orange or green in a matt glaze, it is of great advantage
to use a small' amount of rutile. The rutile appears to "set"
the color and prevent the irregularity of color so often produced
by uranium. An excellent green may be produced by the use of
:obalt and rutile with the addition of a small amount of oxide
of iron, although the color is not quite so bright as that secured
with the presence of some uranium.

The mottling of the color in the matt glaze is undoubtedly due
to the failure of the glaze to dissolve the uranium and the conse-
quent crystallization or segregation, upon cooling. This diffi-
culty is almost always encountered in producing uranium yellow
matt glazes, where reducing conditions may occur, the result
being a mottled green-yellow. The addition of rutile will entirely
overcome this defect and the color will be improved if the con-
tent of zinc oxide is kept very low. A high zinc content and re-
hieing atmosphere in the kiln are favorable for the production
3f greens. The use of the soluble form of the cobalt, rather
than the difficultly soluble stain, as stated by Mr. Wilson, proba-
bly caused the formation of the green color by solution of the
cobalt and uranium instead of by suspension.

HEWITT Wilson: The uranium-cobalt greens developed in
tin above study are not of the same shade as the typical chromium
greens. The color is a deeper green than that afforded by the
majority of the chromium greens burned in terra cotta muffle
kilns to cone 6. While it is true that the cobalt-uranium greens
do not match the characteristic chromium greens, it is likewise

246 AMERICAN CERAMIC SOCIETY.

true that it would be difficult to match some truly acceptable
uranium cobalt green shades with chromium combinations.

There is no question as to the practicability of producing green
colors from cobalt and uranium mixtures. They have been pro-
duced in regular terra COtta work in a quantity sufficient to in-
sure the uniformity required.

The disadvantages of the high cost of uranium is not so great
when the comparatively small amount of green glazed terra
cotta produced is considered. We do not know of any single
component for colored glazes which has caused more terra cotta
trouble than chromium.

The advantages of the above uranium oxide cobalt sulphate
green are as follows :

i. Ease of preparation. No calcining, washing or extra grind
ing necessary.

2. vSo far as noted, there was almost no discoloration of ad-
joining glazes — regardless of their composition.

3. A greater range of glaze composition in which green color!
can be produced. It can be used in zinc and tin glazes.

Although it is not claimed that uranium-cobalt greens will
replace the chromium greens, they have a place in terra cotta
work in which uranium shades are desired and the complexities
of chromium troubles are to be avoided.

Ohio State University,
Columbus, Ohio.

ALABASTER GLASS: HISTORY AND COMPOSITION.

By Alexander Silverman, Pittsburgh, Pa.

Alabaster is a term usually applied to semi-opaque glasses
diich transmit the light from a source, diffusing it, but not ma-
erially altering its color. The name was probably applied be-
ause of the close resemblance which these glasses bear to natural
labaster. The French use the term "Albatre" or "Pate de
tiz" (rice paste). Opal glasses are also semi-opaque but the
ight transmitted acquires an opal or fiery color.

During researches by the writer on the production of alabaster
nd opal effects considerable historical material has been gath-
red and is presented here for ceramists who may be interested.

In the museum exhibits one frequently encounters collections
f Chinese snuff bottles which are made of alabaster glass. Some
f these,1 coming from the Ch'ien Lung dynasty (1736-1795),
.re almost 200 years old. According to one authority, alabaster
;lass was made in Venice and Murano at about 1600. Some
arly Egyptian glasses show slight alabaster effects which are
ffobably due to imperfect melting or to devitrification during
he ages in which they lay buried.

"Manuel Complet de Verrier," by M. Julia de Fontenelle,
>ublished in Paris in 1829, mentions alabaster glass on page 30.

H. Leng2 speaks of the use of lead sulphate in the manufac-
ure of opal glass. The writer's experience, certain details of
chich will be given later, points to alabaster formation in the
>resence of sulphates, so it is possible that Leng's opal was
ilabaster.

Deming Jarvis:i gives the following formula for alabaster glass:
'500 lbs. batch (probably crystal glass), 30 lbs. phosphate of
oda, 10 lbs. allumine (calcined alum), 3 lbs. calcined magnesium."
)n page 116, Jarvis states that alum loses 60 per cent, on cal-

1 Carnegie Museum, Pittsburgh, Pa.
- "Handbuch der Glasfabrikation," 1854.
"Reminiscences of Glass Making," New York, 1865, p. 106.

•is JOURNAL OF THE

cining. Assuming its conversion to .\l.<>: and k:S< ),, it should
lose 70.8 per cent., and it is therefore probable thai the material
was either partly decomposed potash alum or some other sub
stance called alum in that day.

llocii1 mentions alabaster as one of the oldest known glasses
and states that melasse (beet sugar potash bearing sulphati
and chlorides) was employed.

Raimund Gerner2 mentions the use of salt in the preparation

of cloudy glass and gives the following formulas:

1 . 2.

Parts Part*

Sand 100 o Sand 100. 0

Slaked lime 2.0 Slaked lime [3.0

Hone ash 5.5 Tin oxide 1 n

Potash 44 o Potash 38 . o

Salt 5.5 Salt [.5

Arsenious oxide 0.7 Arsenious oxide 0.25

Cullet 50.0 Borax 1 0

It is unlikely that the alabaster effect was produced by the salt
alone. The bone ash probably functioned in batch 1 and the
tin oxide in batch 2, and the arsenious oxide in both.

Before the extensive working of the Stassfurt mines in Ger-
many, much of the potash was derived from plant ashes and
large quantities were imported from Russia. This material
contained considerable quantities of sulphates and chlorides
which were partly conducive to the production of the alabaster
effect.

In the second edition of Gerner's book, Leipzig, 1897, Pa£e 2II>
is given the following batch:

Parts. Parts.

Sand 50 Cryolite 7

Soda 8 Litharge 2

Alumina ( Alaunerde) 4 Lime 3

.Saltpetre 2

On page 34 of the same publication he states that "potash,
until very recently, was prepared from wood and plant ashes.
Beech ashes (Buchenholzasche) contain K2CO3 15 per cent.,

1 Dingler's polytech J., 224, 623 (1877).

2 "Die Glasfabrikation," Vienna, 1880, p. 267.

2co3

K2.S04.

KC1

8

16

3

o

I

9

3

14

2

4

6

6

2

3

3

AMERICAN CERAMIC SOCIETY. 249

Na2C03 3 per cent., K2S04 2 per cent., and the remainder water-
insoluble matter." On page 35 he states that "Refined beet
potash contains K2CO;i 92 per cent., Na2CO;i 5 per cent., KC1
3 per cent., also K2S04." On page 36 he gives the following

analyses of potashes:

R.C03.

American potash 71

Illyrian potash 89

Russian potash 69

Wool Sweat (Suint) 72

Refined Beet Ashes (Melasse) .... 90

Stassfurt 97 0 1 1

On page 59 he dwells upon the use of spars and alumina and
on page 212 states that "alumina may be substituted for the
cryolite, feldspars, etc."

Gessner1 refers to the use of alumina in quantities of from 7
to 20 per cent.

Furnival2 states that American pearlash contains about 7 per
cent. K2SO4.

Hans Schnurpfeil3 also calls attention to the presence of sul-
phates and chlorides in potash.

Fontenelle and Malepeyre1 state that equal parts of chalk
and AI2O3 when melted together yield a glass resembling pearls,
and give a formula for the preparation of a white, translucent
glass, as follows: Sand 100, chalk 32, Al20;i 68. In vol. II, p.
242, they discuss the manufacture of alabaster glass in Bohemia.

Paul Randau5 gives the following alabaster batch containing
sulphates :

Feldspar 35 to 71 parts

100 parts Fluorspar 17 to 50

; Barytes 12 to 40

melted with Soda 15 to 50

or Potash 20 to 65

and vSand 70 to 120

1 Glassmakers Handbook, Pittsburgh, 1891, p. 148.
- "Researches on Leadless Glazes," 1898, p. 54.

3 "Schmelzung der Glaser," 1906, p. 11.

4 "Manuel du Yerrier," Paris, 1, 59 (1900).

"Die Farbigen, Bunten and Verzierten Glaser," Vienna, 1905, p. 106.

2.so JOURNAL OF THE

These proportions arc also given in a patent1 granted to J.

Kemper.

An article by Richard Zsigmondy* on "Cryolite and its Sub-
stitutes," mentions the use of china clay and sodium fluoride
in alabaster batches, stating that the best results are obtained
by adding a quantity of china clay equivalent to the sodium
fluoride in the cryolite.

The following formulas11 are suggested for the use of alumina
and sodium fluoride:

Parts. Parts.

Sand ioo ioo

vSodium fluoride 10 6

Alumina (Alaunerde) 14 10

Fluorspar o 10

Feldspar o 8

The use of sodium fluoride in alabaster glass is covered in a
patent4 issued to Adolf Tedesco, who gives the following formula

Parts.

Sand 156

Sodium fluoride 33 — -(containing 10 per cent, sodium carbonate)

China clay 15

Chalk 10

Soda 5

In this patent the inventor claims the use of sodium fluoride
only, so that the use of alumina in alabaster batches must hav(
been known to glass makers.

Weinreb5 mentions the use of a mixture of Si02 100 parts
Na2F2 20 parts, K2C03 8 parts, Na2C03 7 parts, CaC03 8 parts
and A12(OH)g 6 parts, which, he states, yielded a beautiful, whit
glass.

In the Journal of the Society of Chemical Industry6 is note
a statement that "Since, however, fluorine (from cryolite) al
tacks the glass-pots and thus leads to contamination of the glas^

1 D. R. P. 4551, July, 1878, Dingier' 's polytech. J., 233, 269 (1879).

2 Ibid., 271, 40 (1889).

3 Sprechsaal, 30, 556 (1897).

4 D. R. P., 31,112, November, 1883. .

5 Dingier' s polytech. J., 256, 365 (1885).

6 /. Soc. Chem. Ind., 18, 916 (1899).

AMERICAN CERAMIC SOCIETY. 251

Idspar or alumina is sometimes added, the composition of the
Lass being modified as follows: Sand 50 parts, potash 4 or 5
arts, soda ash 5 or 6 parts, cryolite 7 parts, feldspar or alumina

parts, lime 5 parts, and nitre 3 parts. A cheap substitute for
yolite is fluorspar which, however, has to be neutralized by
umina in order to prevent corrosion of the pots, etc. A typical
lass is prepared from a mixture of sand 50 parts, soda ash 8 to
d parts, potash 3 to 4 parts, fluorspar 10 parts, alumina 5 to 10
arts, nitre 2 to 3 parts, zinc white 2 to 3 parts, feldspar o to 10
arts, etc. A mixture of sodium fluoride and alumina is also
leaper than cryolite, etc. The mass is compounded from a
ixture of sand 50 parts, potash 6 parts, soda ash. 6 parts,
idium fluoride 4 to 6 parts, alumina 5 to 8 parts, feldspar 6
arts, nitre 1 part, minimum 3 parts, or zinc white 2 parts, and
me o to 3 parts."

The references cited in the preceding paragraphs show clearly
lat the use of mixtures of alumina, sulphates and chlorides
ith other raw materials is not original with American chemists
id manufacturers although they have produced some of the
nest alabaster glasses of superior quality and commercial value,
o explanation of the cause of the alabaster effects is offered by the
irly writers except that of devitrification. The author believes
iat sulphates and chlorides serve as ionized electrolytes in the
3t liquid, precipitating various colloidal suspensions which would
:herwise cause opalescence.1

In the writer's researches on alabaster glass a number of in-
resting facts have been developed. Fluorides with alumina or
uminium bearing substances produce opalescence until suffi-
ent quantities of sulphates or chlorides are introduced when
le opacifying materials gather in larger particles and white
*ht is transmitted. The bivalent ion of the sulphates produces
.ore intense whiteness than the monovalent chloride. This is
i keeping with the behavior of electrolytes towards aqueous
)lloidal suspensions.

The presence of barium2 compounds enhances the white color —
1 A. Silverman, "Similarity of Vitreous and Aqueous Solutions," /. Ind .
ng. Chem., 9, 33 (1917)-

-A. Silverman, "The Use of Barium Compounds in Glass," J. Soc.
hem ln,l., 34, 399 (1915).

Ji >i RNAL OF THE

probably through the formation of barium fluosilicate. I
produces increasing whiteness up to a certain concentratioi
The color tlieu gradually loses intensity wit h the introductu
of larger quantities and finally disappears. This effect
bles the solution of precipitates in aqueous solutions in the pre
ence of an excess of the precipitant.

Alumina and the chlorides were employed in the writer's n
searches prior to 1904 and sulphates were introduced in 191
As the result of these researches a commercially successful al<
baster glass1 of American manufacture was placed on the marke
The limited solubility of the chlorides and sulphates in the sil
cates led to a considerable reduction of the quantities introduce
into the batch, with an accompanying increase in the life of tl
pots which are attacked by fused chlorides and sulphates.

When the alumina was replaced by its equivalent of feldspa
a beautiful, speckless glass of much higher than normal viscosit
resulted. Magnesium silicate also produced the alabaster e
feet. The following formulas for American alabaster glasse
are contained in patents:2

Parts.

IOO

I55A

2I7/8

Pad

Sand

Lead oxide

Soda

Niter

Salt

Borax

Aluminium hydroxide .
Fluorspar

Sand 34J

5 1

sV«

I1/4t0 2>/,
I8.I2

6

Litharge

Soda ash

Cryolite

Aluminium oxide.

Niter

Borax

Plaster of paris. . .

Parts.

IOO

56

36
20
60

Sand

.Soda ash 49

Red lead 56

Feldspar 120

Sodium fluoride 71

Strontium sulphate 1 3

Sodium chloride 5

Alumina 2 2 '

Xiter

Antimonious oxide

A. Silverman, "The Chemist and the Glass Manufacturer,'
Am. Ceram. Soc, 12, 187 (1910).

: U. S. Patents 1,097,600-1,143,788 and 1,245,487.

1 „

Tra

AMERICAN CERAMIC SOCIETY. 253

Although the most desirable alabaster glasses are produced
f aluminium-bearing substances with fluorides and chlorides
• sulphates, this article would not approximate completeness
ithout mention of other types given in the literature. Calcium
losphate, magnesium silicate, stannic oxide, arsenious oxide,
rconium oxide' and titanium oxide2 are also employed. Anal-
;es of some alumina-bearing alabaster glasses have revealed
.e presence of considerable quantities of zirconia, presumably
troduced as an impurity in the alumina.

Alabaster effects may also be obtained in high silica glasses
rough devitrification. Raimund Gerner3 suggests the follow-
er batches :

1. 2. 3. 4.

Sand 200 200 250 100 parts

Pearlash 80 70 100 42

Bone ash 12 to 15

Magnesium silicate 9 3 14 8

Arsenious oxide 1 1

Saltpetre 10 15 5 "

Wilhelm Mertens suggests:4

1. 2. 3.

Sand 75 75 100 parts

Pearlash 28 30 38

Bone ash 7 7

Saltpetre 6 8 4 "

Magnesium silicate 10 10 "

Red lead . . 12 "

Arsenious oxide 1 0.5 . . "

Randau5 gives this batch: Sand 100 parts, calcium car-
mate 40 parts, talc 5 parts and borax 5 parts
Rudolf Hohlbaum6 suggests:

1 D. R. P., 189,134.
'-' D. R. p., 18,708.

3 "Die Glasf abdication," p. 206.

4 "Die Fabrikation und Raffinierung des Glases," p. 214.

' "Die Farbigen, Bunten und Verzierten Glaser," 1905, p. 112.
1 "Herstellung, Bearbeitung und Verzierung des Feineren Hohlglases,"
10, p. 135.

254

JOURNAL OF THE

Sand Km leg.

Pearlasfa (65* ,)....

Calcite

Hi )iu ash

Magnesium silicate.

Saltpetre

Manganese dioxide.

41 leg.

IO kg.

200 g.

2

3.

roo

kg.

IOO kg

(80

to

85

!

4.5
17
1 1

200

kg.
kg.
kg.

g.

60 to 65%)
Guano

40 kg

7 kg

2 kg

3 kg
200 g.

Sand 1 00 parts

Pearl ash (60 to 65%)

Red lead

Boric acid

vSaltpetre

Bone ash

Stannic oxide

39
30

7
4

Additional formulas are mentioned in other books and ii
journals. Most of these batches fail to work in American fac-
tories. They are usually intended for low temperature slow
melting furnaces. In some foreign factories the glass is ladlec
into water and then remelted.

It may now prove of interest to consider the break of continuit
in the history of alabaster glass caused by the introduction
opal glass. The earlier white glasses were chiefly alabaster, du
as has already been intimated, to the effect of chlorides and si

AMERICAN CERAMIC SOCIETY. 255

phates, present as impurities in the various types of potash,
and causing the precipitation of the colloids in the glass. When
the purer refined pearl ash, prepared from Stassfurt salts and
sugar beet waste, became available the absence of powerful
electrolytes, or the presence of only very small quantities of
electrolytes, prevented precipitation and the smoother opal
glasses came into general use. During the latter decades of the
nineteenth century opal glass displaced albaster to a great ex-
tent. When American manufacturers succeeded in preparing
it on a commercial scale in the early part of the present century
they felt that they had made a discovery. What they really
bad done was to introduce the chloride or sulphate, missing in the
pure pearl ash, and so necessary for precipitation. In other
words, they obtained through the application of scientific princi-
ples a result which impurities in their raw materials had acci-
dentally given to the foreign manufacturers.

The writer conducted an extensive correspondence with glass
makers in England, France, Belgium, Italy, Germany and Aus-
tria during 19 13 in order to secure definite historical data' in
reference to alabaster glass. The diversity of opinion which
follows shows how difficult it is to settle such questions. The
date indicates the year in which the manufacture of alabaster
glass began according to the informant whose name is given.
Where the date is marked (S) a specimen was received which is
in the writer's collection. (M) indicates that the informant
manufactured alabaster glass. A photograph of some of these
specimens is shown on the preceding page.

Manufacturers of Alabaster Glass.
1600 The Venice & Murano Co., Venice, Italy.

1700 Josef Janke & Co., Haida, Austria. (Manufactured by Chinese

and Japanese.)
1 8 10 Josef Knizek, Ullersdorf, Austria.

1 810 E. Michel & Co., Teichstadt, Austria.

1810 Josef Rindskopf's Soehne, Teplitz-Schoenau, Austria.

1810 J. Schreiber & Neffen, Wien, Austria.

1820 Josef Janke & Co., Teichstadt, Austria. (1820 Haida, Bohemia.

After 1840 in nearly all Bohemian huts. Made with Russian

potash instead of ordinary potash. Many old specimens in

Janke museum.)

JOURNAL I '!•' THE
i x jo [osef Rindkopf's Soehne, Dux, Austria. (Made by father in

Tischen, Austria, iMin

M Wilhelm Kralik, Johann Mayer Factory, Winterberg, Au
S 1840 Compagnie des Cristallerie de Baccarat, Paris, Prance.

1840 M Meyer & NTenen, Winterberg, Austria.

0 M Brueder Gerhart, Mulan and Podiebrad, Austria.
S 1850 Jilck & Vetter, Steinschoenau, Austria. Colored lis- bone ash,
talc and magnesium silicate.
[850 M Carl Goldberg, Haida, Austria. Has piece manufactured by

rather in [850.
[850 M Arthur Rudd cN; Co., Lancashire-, England. (Also made at

Birmingham, Stourbridge and London.;
1850 M Molineux Webb & Co., Ltd., Manchester, England. (Trade name
"Carnelian.")
S [860 M E. Michel & Co., Teichstatt, Austria.

S 1863 Herman Hcye, Hamburg, Germany. (Bowl in illustration.)

S 1864 M Bartles, Tate & Co., Manchester, England. (Trade mark "'Cor-
nelian." )
S i860 M Raimund Knospel and .Soehne, Hillemuehl, Austria.

i860 M E. St. Clair (Crystallerie de Baccarat), London, England. (Trade

Mark "Agate.")
1870 M J. Schreiber and Neffen, Gr. Ullersdorf, Austria.
S 1870 M Tietze and Seidensticker, Penzig, Germany.
1870 M Constantin Kopp, Klein-Anjezd, Austria.

1880 M August Walter Soehne, Moritzdorf, Germany. (Vases and il-
luminating ware.)
1880 M Josef Palme, Post Kitlitz, Austria.
1880 M Glashuetten-Werke Carlsfeld, Carlsfeld, Germany.
S 1884 M Anton Rueckel & Sons, Nischberg, Austria.

S 1888 M Joseph Knizek, Ullersdorf, Austria. (Small vase in illustration
made by father. Batch: sand 150, potash 65, magnesium
silicate 7, saltpetre 6, bone ash 8.)
1890 M Wilhelm Kralik Sohn, Eleonarehhain, Austria.
S 1892 M Val Saint Lambert Glassworks, Belgium. (Lamp fount in il-
lustration.)
1898 M Glasindustrie Schreiber Aktiengesellschaft, Berlin, Germany.

Bibliography — Books.
Agricola, De la Fabrication du Verre, 1612.
Asch, W., and Asch, D., The Silicates of Chemistry and Commerce, New

York, 1914.
Benrath, H. E-, Die Glasfabrikation, Vienna, 1874.
Biser, B. F., Elements of Glass and Glassmaking, Pittsburgh, 1900.
Bontemps, G., Guide de Verrier, Paris, 1868.
Dammer, O., Chemische Technologie der Neuzeit, v. 3, Stuttgart, 1910.

AMERICAN CERAMIC SOCIETY. 257

alk-, R., Die Glasfabrikation, Munich, [911.

card, J., Le Verre et sa Fabrikation en Four Electrique, Paris, 1907.

ntenelle et Malpeyre, Manual du Verrier, Paris, 1900.

ntenelle, M. J., de Manuel Complet de Verrier, Paris, 1829.

nnm, Le Verrerie du XIXe Siecle, Paris, 1863.

irnival, W. J., Researches on Leadless Glazes, 1898.

undwald, J., and Hodgson, H. H., Raw Materials of the Enamel Industry,

London, 1914.

issuer, F. M., Glass Makers Handbook, Pittsburgh, 1891.

Slier, R., Die Glasfabrikation, Vienna, 1897.

ludiequier de Blancourt, J., de l'art de la Verrerie, Paris, 1697.

)hlbaum, R., Zeitgemasse Herstellung, Bearbeitung und Verzierung des

Feineren Hohlglases, Vienna, 19 10.

mrivaux, Kncyclopedie Chimique, Paris, 1883.

mrivaux, J. L. C., La Verrerie au XXe Siecle, Paris, 1903.

■nrivaux, J., La Verrerie, Paris, 191 1.

jvestadt, H., Jena Glass and its Scientific and Industrial Applications,

London, 1902.

rvis, D., Reminiscences of Glass Making, New York, 1865.

'ram Rundshau, Taschenbuch fur Keramiker.

inge, G., Technical Methods of Chemical Analysis, v. 1, pt. 2, London,

1908.

1 Chatelier, H., La Silice et les Silicates, Paris, 1914.

Bg, H., Handbuch der Glasfabrikation, 1854.

ellor, J. W., A Treatise on Chemical Analysis, v. 1, London, 1913.

inutoli, De la Fabrication et de l'Emploi des Verres colores chez les Romains,

Berlin, 1836.

ertens, W., Die Glasfabrikation, Vienna.

uller, B., Chemische Technologie des Glases, Leipzig, 1911.

isbitt, A., Glass, New York, 1879.

•well, H. J., Principles of Glass Making, London, 1883.

ligot, K., La Verre, son Histoire, sa Fabrikation, Paris, 1878.

>senhain, W.( Glass Manufacture, New York, 1908.

indau, P., Fabrikation der Farbigen, Bunten und Verzierten Glaser,

Vienna, 1905.

irechsaal Kalender.

hnurpfeil, H., Two Hundred Recipes for Glasses of All Kinds, Zurich, 1918.

hnurpfeil, H., Schmelzung der Glaser, Vienna, 1906.

hnurpfeil, P., Das Glas in Seiner Verschiedenen Farbentechnik, Strassburg,

191 1.

.uzay, La Verrerie depuis les temps les plus Recules jusqu'a nos Jours,

Paris, 1868.

ein, Fabrikation du Verre, Brunswick, 1862.

imbon, A., Les Verres Antique.

;euschner, K., Glasfabrikation, Weimar, 1885.

JOURNAL OF THE
in, Dictionary <>i Aits, Manufacturers and Mines, v. 2, Ixmdon, 1875.

Uhlig, B. C, Chemical Analysis for C.lassniakcrs, Pittsburgh, 1003.

Zschimmer, E., Die Glasindustrie in Jena, 1909.

Bibliography — Journal Articles.

Alabaster Glass.
Alabaster glass, W. Stein, Jahrb. Ber., 1, 151-52 (1855); 4, 245 (1858).
Alabaster glass, Bull. soc. chim.. A, I, 376 (1858).
Alabaster glass, W. Stein, Dingier' s polytech. J., 152, 75-76 (1859).
Alabaster glass, Examination of, W. Stein, Chem. Zenlr., 30, 207 (1859).
Alabaster glass, Jahrb. Ber., 16, 279-280 (1870).
Alabaster glass, Analysis of, Stolba, Dingier' s polytech. J., 198, 178-179 (1870I
Alabaster glass, M. Hock, Ibid. , 224, 623-28 (1877).
Alabaster and opal glass, Jahrb. Ber., 23, 499-500 (1887).
Alabaster glass, Analysis of, Chem. Zentr., 51, 76 (1880).
Alabaster glass, Jahrb. Ber., 37, 356 (1891).
Alabaster glass, Ibid., 39, 691-92 (1893).

Alabaster glass, Sprechsaal, 28, 389 (1895J; Jahrb. Ber., 41, 733-34 (1895)
Alabaster glass, Jahrb. Ber., 43, 710 (1897;.
Alabaster glass, Sprechsaal, 30, 497 (1897).
Alabaster and opal glass, Ibid., 35, 3-4 (1902).
Alabaster glass, Ibid., 43, 321 (1910J.
Alabaster glass, cast, yellow spots, Ibid., 45, 95 (1912).
Alabaster glass, G. A. Macbeth, U. S. Pat. 1,097,600 (1914).
Alabaster Glass, Use of, in production of "Mother of Pearl" glass, Sprechs*

48, 293 (1915)-
Alabaster glass, H. S. Schnelbach, U. S. Pat. 1,143,788 (1915).
Alabaster glass, J. J. Miller, U. S. Pat, 1,245,487 (1917).

Alumina in Glass.
Alumina glass, H. Reinsch, Jahrb. Ber., 6, 293 (i860).
Alumina glass, J. Pelouze, Ibid., 11, 424 (1865).
Alumina glass, J. Pelouze, Ann. chim. phys., 245, 468 (1865).
Alumina glass, J. Pelouze, Bull soc. chim., n. s., 7, 458-60 (1870).
Alumina glass, Jahrb. Ber., 13, 349-51 (1867 J.
Alumina glass, J. Pelouze, Ann. chim. phys., 250, 188 (1867).
Alumina glass, J. Pelouze, J. prakt. Chem., 101, 449, 452-53 (1867).
Slag, Utilization of, for the manufacture of glass, Britten, /. Iron Steel Inst.

10, 453-66 (1876J.
Alumina in glass, Jahrb. Ber., 34, 775-80 (1880).
Alumina glass, Ibid., 27, 459 (1881).
Alumina glass, Preparation of high alumina glass from volcanic silicates

Ber., 15, 265 (1882).
Alumina, Influence of, Jahrb. Ber., 31, 575 (1885).
Alumina, Influence of, Verh. Ver. Berford Gewerb, 66, 799-800 (1887).

AMERICAN CERAMIC SOCIETY. 259

Alumina, glass with high content, Frank, Ibid., 67, 406-7 (1888).

Alumina in glass, Henrivaux, Chem. Zentr., 60, 850 (1889).

Alumina, Influence of, Weber, Verh. Ver. Berford Cewerb, 68, 188-94 (1889).

Alumina in glass, Jahrb. Ber., 38, 613, 617, 618, 623, 627, 632 (1892).

Alumina glass, Ibid., 39, 690 (1893).

Aluminium borate, Use in glass, Ibid., 39, 687 (1893).

Alumina in glass, L. Appert, Compt. rend., 122, 672-73 (1896).

Alumina glass, Appert, Jahrb. Ber., 42, 688 (1896).

Alumina, Effects of, on glass, R. L. Frink, Trans. Am. Ceram. Soc, II, 99-100

(1909).
Alumina in glass, Sprechsaal, 43, 118 (1910).
Alumina in glass, F. Gelstharp, Trans. Am. Ceram. Soc, 12, 329, 331, 821

(1910).
Alumina, Effect of, in glass, F. Gelstharp, Ibid., 12, 330 (1910).
Alumina, Influence of, on color in glass. Ibid., 13, 256 (191 1).
Alumina, Use of, in glass, Ibid., 14, 364 (1912).
Alumina, Use of, in glass, Ibid., 14, 70 (191 2).
Alumina in glasses, A. Granger, Mon. Sci., 82, 274-85 (1915).
Alumina, Influence of, on the fusibility of glasses, L. Springer, Keram. Rund-
schau, 23, 271-2, 280 (1915).
Alumina, Influence of, on fusibility of glasses, F. Singer, Ibid., 23, 65-6, 78-9,

95-6, 103-4 (1915)-

Chlorides and Sulphates in Glass.
Sodium sulphate, Use of, in glass making de Serres, Ann. chim. phys., 76,

172-85 (1810)
Sodium sulphate, Use in glass making, Gehlen, Ibid., 97, 216-17 (18 16).
Salt and Sodium sulphate, Glass made with, Leguay, Dingier' 's polytech. J.,

18, 235-36 (1825).
Sulphate content, Jahrb. Ber., 11, 424-425 (1865).
Soldium sulphate in glass making, Chem. News, 15, 159 (1867).
Sodium sulphate, Use in glass making, R. Wagner, Bull. soc. chim., n. s., 23,

379 (1875)-
Sodium sulphate, Use in glass making, O. Schott, Dingier' s polytech. J., 221,

142-46 (1876).
Salt, Use in glass making, Jahrb. Ber., 23, 472-74 (1877).
Potash, Composition of, M. Hock, Dingler's polytech. J., 224, 623 (1877).
Sodium Salts, Action of, Jahrb. Ber., 38, 622-26 (1892).

Sulphate glass, Manufacture of, Goerisch, J. Soc. Chem. Ind., 16, 680 (1897).
Sulphate glass, Jahrb. Ber., 45, 673 (1899).
Sulphate glass, Ibid., 49, 407 (1903).
Sodium sulphate, State in glass, Bull. soc. chim., 95, 379 (1908); Z. angew.

Chem., 20, 1899-1900 (1907).
Sulphates in glass, Trans. Am. Ceram. Soc, 12, 330, 334 (1910).
Salt and sodium sulphate, Solubility of, in glass, F. Gelstharp, Ibid., 14,

665-67 (1912).

260 Ji lURNAI ( >F THE

Sodium sulphate, I se of, as a batch ingredient, Sprechsaal, 49, 38, 286 (1916).
Chlorides and sulphates, Influence of, in producing opalescence in
J. D. Cawood and W. K. S. Turner, ./. Soc. Glass Tech-, 1. 87 93 (191

Similarity of aqueous and vitreous solutions, A. Silvermann, ./. Ind
Chem., g, 33 1 1017).

Opal Glass.
Opal glass, Jahrb. Her., 23, 499-503 (1877).
( >pal glass, Weinreb, Dingler's poly tech. J, 256, 361 -67 (1885).
Opal glass, Hock, Ibid., 224, 623^28 C1877;.

Opal glass, toughened, C. D. Abel, J. Soc. Chem. Ind., 4, 535 (1885).
Opal glass, A. Tedesco, Dingler's polytech. J., 271, 425 (1889).
Opal glass, by silicon hydrofluoride, G. Cesaro, Ber., [4] 27, 335 (1894); Bull.

Acad. Roy de Belgique, [3] 26, 721-30.
Opal glass, F. H., Sprechsaal, 29, 155, 185, 213, 243, 271, 299 (1896).
Opal glass, W. M., [hid., 30, 556 (1897).
Opal glass, Constitution of, Jahrb. Ber., 45, 677-79 (1899).
Opal and alabaster glass, Sprechsaal, 35, 3-4 (1902).

Opal glass, translucent, J. Kempner, /. Soc. Chem. Ind., 24, 133 (1905).
Opal glass from blast furnace slag, Le Chatelier, Ibid., 24, 970 (1905).
Opal glass, H. Schnurpfeil, Die Glashutte, 38, 380-93 (1907).
Opal glass, Pure white, Sprechsaal, 45, 96 (1912).
Opal glass for tiles, Ibid., 46, 76 (1913).
Opal glass, Watery, Ibid., 46, 608 (1913).
Opal glass in optical glass, O. Bloch, and F. F. Renwick, Phot. J., 56, 49-66

(1916).
Opal glass, Difficulties in' the manufacture of lamp shades, Sprechsaal, 46,

6, 47 (1916); J. Soc. Glass Tech. (Abstract Section), I, 11-12 (1917).

Opaque Glass.
Opaque glass, Bull. soc. chim., A, V, 152.
Opaque glass, Lead sulphate as substitute for tin oxide, Dingler's polytech. J.,

127, 464 (1853).
Opaque glass, /. Soc. Arts, 26, 161 (1878).

Opaque or opalescent glass, J. H. Johnson, /. Soc. Chem. Ind., 4, 349 (1885)
Opaque glass, Jahrb. Ber., 38, 635 (1892).

Opaque glass, W. Hirsch and A. Tedesco, Chem. Zentr., [2] 64, 896 (1893).
Opaque glass, W. Hirsch and A. Tedesco, Dingler's polytech. J., 297, 282 (1895)
Opaque glass, Diamant, 21, 408, 429-30 (1899); J. Soc. Chem. Ind., 18, 916

(1899).
Opaque colored glass, Sprechsaal, 32, 322-24, 352-53 (1899J; /. Soc. Chem.

Ind., 18, 917 (1899).
Opaque glass, H. E. Knospel, Ibid., 21, 347 (1902).
Opaque glass, R. Rickmann & E. Rappe, Ibid., 22, 544 (1903).
Opaque glass, Manufacture of, G. E. Barton and A. V. Bleininger, Die

Glashutte, 37, 210, 226, 237 (1907).

AMERICAN CERAMIC SOCIETY. 261

Dpaquc glass, Reinmann, /. Soc. Client. Ind., 26, 610 (1907).

Dpaque to active light with silver, O. Sackur, Chem. Zentr., [1, B] 79, 1817

(1908).
Dpaque glass to actinic light, O. Sackur, /. Soc. Chem. Ind., 27, 567 (1908).
Dpaque glass, Chemistry of, E- Knequist, Chem. Eng., 10, 54-55, 128-129

(1909).
Dpaque semitranslucent glass, J. L- Miller, U. S. Pat. 1,245,487, Nov. 6, 1917.

Opalescent Glass.
Dpalescent glass, O. Reinsch, Dingier's polytech. J., 184, 369-73 (1867).
Dpacity in glass, P. Weiskopf, Ibid., 206, 468 (1872).
Dpalescent glass. C. Huelser, J". Soc. Chem. Ind., 10, 139 (1891).
Dpalescent glass, Process for making, Bd. of Trade J., Nov., 1899, 618 (1899);

/. Soc. Chem. Ind., 18, T067 (1899).
Dpalescent glass, American, Alexander, Sprechsaal, 36, 78 (1903).
Dpalescent glass, process for making, A. Lesmiiller, J. Soc. Chem. Ind., 29,

426 (1910).
Dpalescence in glass, Causes of, J. G. Smull, Ibid., 34 402-405 (1915).
Dpalescence in glass, Influence of chlorides and sulphates in producing, /.

Soc. Glass. Tech., 1, 87-96 (1917).

(See Sprechsaal, Keram Rundschau, Glas-Indtislrie and Diamant, complete
ets of which were not available in Pittsburgh when this list was prepared.)
School of Chemistry,

I'NiVERSITY OF PlTTSBl'ROH.

THE ZWERMANN TUNNEL KILN AND ITS OPERATION.

By Carl H. Zwkkmann. Kalamazoo, Mich

It would appear that the following points are essential in tin-
construction and operation of a successful tunnel kiln:

First, as the tunnel kiln, on account of its fuel and labor-
saving qualities, is intended to replace the periodic kiln, it would
appear that it must be so constructed and operated as to enable
one to produce the same conditions as in a periodic kiln.

Second, the succees of a tunnel kiln depends more upon hav-
ing it under good control than upon any other factor.

Third, as the first cost of a tunnel kiln is considerably higher
than that of a number of periodic kilns of equal capacity, it
should be so constructed as to have a comparatively wide tunnel.
A tunnel kiln twelve feet wide costs comparatively little more
than a tunnel kiln four feet wide, and has about three times the
capacity — the heat loss through radiation being proportionately
less.

Fourth, the design of the kiln should be as simple as possi-
ble so as to require the minimum of repairs. It must be so con-
structed as to reduce to a minimum the heat losses through radia-
tion and so as to utilize the heat from the cooling zone for drying
or heating purposes.

The Zwermann oil-fired kiln, illustrated in Fig. i, is being
operated at Kalamazoo. This kiln is about 352 feet long, in-
cluding the vestibule, the tunnel being 6' 4" wide and 6' 6" high
from the top of the trucks to the arch. It will hold 46 trucks,
including the one in the vestibule, each truck being 7' 71, 2" long.
There are nineteen trucks in the preheating zone up to the first
burner, eight trucks in the firing zone and eighteen trucks in
the cooling zone.

The entrance to the kiln is through a vestibule provided with
two tightly fitting rolling doors. A rolling door was originally
provided at the outlet end of the kiln but this has been found un-
necessary. A hydraulic pusher is installed at the entrance to

£*haaster

- Air inlet in ar<rh

Discharging Cnd

3/atver
• Air Inlet

5ect,onA-A Section B-B SectionC-C Section D-D Section H Sect/on f-f SectionC-G Sectionlt-H Section J-J Section^

Zi/t~/?AMm5 TrnmriA/iA1

AMERICAN CERAMIC SOCIETY. 263

te kiln, the pusher rod passing through an opening in the vesti-
jle door, the entire train being advanced one truck length

a time. About three minutes are required to introduce and
move one truck. The loaded trucks are introduced at two-
>ur intervals, 92 hours being required to pass a truck through
le kiln.

The kiln gases are drawn off by means of an exhaust fan near
ic entrance end. The waste gas outlets are at the fourth truck
0111 the vestibule, on a level with the tops of the trucks, and
ad through a flue in the sides of the kiln, extending as far as
lick No. i and then out through the exhaust. Mechanical
raft was preferred as it permits of closer control than stack
raft. A second exhaust, located about midway in the length
: the kiln, draws cold air through the hollow arch and side
alls of the cooling zone of the kiln and passes it through the
are. This pre-heated air is utilized in the drying and pre-
dating of the ware. The ware is indirectly cooled by this air,
hich does not pass through the tunnel or come into contact
ith the ware. The heat from the cooling ware in both kilns is
lfhcient to heat the factory and dryers during very severe
eather.

The pressure blower, which delivers the air for combustion
> the burners, is located on the outlet end of the kiln. This
r is passed through recuperators built in the side walls of the
moling zone.

Tlu trucks are of substantial steel construction with cast iron
hills and have a. plate on either side which runs in a sand seal,
ach truck is covered with a layer of insulating material, on
)p of which is placed a layer of fire brick tiles. The ware is
acked around and above a combustion chamber which is built
a the center of each truck. This combustion chamber is at
ght angles to the tunnel.

The kiln walls are constructed from a layer of </' lire brick,
1 •_>" of common brick and a variable thickness of kieselguhr.
Ii is built of 9" lire brick and is also insulated with kiesel-
uhr. The insulation of a tunnel kiln is very necessary in order
) reduce the heat losses through radiation. Temperature
ladings, taken at a point at which the internal temperature of

264 JOURNAL OF THE

the kiln was [43O0 F., gave" ["J ''• at the surface of the outei
wall and 1200 F. at the surface of the top of the arch. At an
nt her point, at which the internal temperature was 23000 F.
the outside surface temperature was [350 F. and the surface
temperature of the top of the arch was 1420 F.

The ten oil burners used are located on both sides of the kiln
At the two last trucks in the firing zone the burners are opposite
each other, the balance being located alternately on either side
of the kiln. Eight trucks, in all, are opposite the burners. Th<
location of the trucks in the kiln is so arranged that when a
rest the center of each combustion chamber on a truck is oppo
site the center of a burner.

The oil is delivered to the burners under pressure and is atom
ized with compressed air. The balance of the air for combus
tion is supplied, pre-heated, by the pressure blower as already
described. All of the air for combustion is blown through th<
burner blocks and excellent combustion is continually main
tained in the combustion chamber. With the arrangement 0
the burners and air supply described, the kiln may be operatee
either with a reducing or oxidizing fire by a slight change of th
dampers controlling the air supply. This is very necessary ii
the burning of certain classes of products. In a periodic kib
the atmosphere is somewhat reducing after every coaling am
oxidizing after the fires have burned down. The same condi
tion may be created in the Zwermann kiln with little or no cliffi
culty.

The combustion chamber, being built on the truck proper, th
making of repairs is simplified and the heat is delivered direct!
to the center and at the bottom of the material being burned
This, and the fact that the kiln is under excellent control, a.c
sures an even distribution of heat and a very uniform firing c
the ware. Furthermore, the location of the combustion cham;
bers on the trucks and transversely to the tunnel permits th I
building of a much wider kiln.

For wares burned in saggers, a- combustion chamber on th
top of each truck is required in order to permit of the ware bein
stacked. If bricks are being burned no special combustion chan

AMERICAN CERAMIC SOCIETY. 265

it structure is necessary — the brick being set so as to form their
A^n combustion chamber.

The burners may be regulated at will and the desired tem-
irature maintained at each one. Eight thermocouples are in-
rted through the arch in the heating zone of the kiln at Kala-
azoo. These are connected to a recording galvanometer.
y providing the necessary sight holes in the burning zone
ie burning may be regulated by means of cones — -although
>tter control is had by the use of pyrometers. The
isired temperature is maintained for ^ays and weeks without
ie pyrometer charts showing large variations. The tempera-
ire of the ware when it reaches the first burner is about 1000° F.
fter the temperature desired at each burner has been determined

can be maintained without great variation.
The waste gases leave the kiln at a temperature of about 300 ° F.
bviously, the percentage of excess air introduced into the kiln
controlled by the regulation of the exhaust.
The quantity of the ware loaded on the cars has been found to
i an unimportant factor in the operation of this kiln. When
Le kiln was first started, there was not enough ware on hand
id some empty trucks were passed through. Those following
ere loaded only to one-half of the height of the tunnel. In
rery instance, however, the ware was well burned, no over-
jrned or under-burned pieces being noticed. This was proba-
y due to the location of the combustion chambers at the top

the trucks as well as to the method of firing and regulating
ie kiln.

We have determined that if the pressure blower, supplying
Le outside air and also the preheated air from the cooling zone
r combustion, was cut out, a large percentage of defective
are and uneven firing resulted. A difference of as high as two
»nes, between the top and bottom of the truck, was noticed,
he highest temperature was at the bottom, although when the
In was operated under proper control the greatest difference

temperature found between the top and bottom of a truck
ad was about one-half cone — -the ware being burned to cone 9.
The first kiln erected at Kalamazoo is 352 feet long; the second
In is 285 feet long. Although a longer kiln permits the for-

266 AMERICAN CERAMIC SOCIETY.

ward movement of the trucks at shorter intervals, the lenfl
■ ■I a kiln should obviously be dependent upon the- capaffl
desired and the material to be burned.

Summary.

The experience gained during the period in which the Kal
mazoo kiln has been operated demonstrates that, iu order 1
produce marketable ware of any kind and particularly sanitl
porcelain, an absolute and convenient control of the kiln is nece
sarv in order that the human factor may be eliminated as far t
possible.

It is our conviction that the Zwermann kiln may be used i
place of the periodic kilns in the firing of many kinds of eeram
wares. The water-smoking and pre-heating zones are und
better and more even control than in a periodic kiln. The coolil
of the ware is done just as efficiently as in a periodic kiln. By tl
use of the tunnel kiln we have made a considerable saving in fli
over that required in the firing of the periodic kilns. A consi<
erable saving in labor has also been noted. A skilled firerru
is not required. As the trucks are loaded and unloaded in tl
open, this work is also lighter and easier and does not requl
the services of skilled kiln setters.

IE PROPERTIES OF SOME OHIO AND PENNSYLVANIA
STONEWARE CLAYS.1

By H. G. SCHURECHT, Columbus, Ohio.

Introduction.

The resistance of vitrified clay to chemical action, together
th its impenetrability to liquids, has made chemical stone-
ire an important product in the chemical industries. Stone-
ire distilling kettles, receivers, filters, condensing worms,
ltrifugal pumps, acid-proof pipe fittings, tower packing, acid-
aof tanks, photographic tanks, and troughs and tubes for elec-
cal purposes are used in these industries. Stoneware equip-
Brt is also used in the various food industries, including the
gar industry. Owing to the increased production of explo-
res during the war, there has been a marked increase in the de-
ind for chemical stoneware and for stoneware clays. The
der use of chemical stoneware is only limited by its compara-
-e brittleness and its sensitiveness to sudden temperature
anges.

Composition and Properties of Chemical Stoneware.

The clays or bodies used in the manufacture of chemical stone-
ire should have good bonding power, a small drying and burn-
l shrinkage, and a somewhat wide vitrification range. Crushed
meware is sometimes added to reduce the shrinkage. Feld-
a.r or a mixture of feldspar and a small percentage of calcium
rbonate is sometimes added as a flux. Since the color of chem-
1 stoneware is not an important consideration, cheaper fluxes,
:h as phonalite, trachyte, porphyry, and retinite, have been
.'(1 in Europe2 in place of feldspar. The chemical analysis of a
rietv of phonalite is as follows:

1 By permission of the Director, Bureau of Mines.

2 Kerl, B., Cramer, E., and Hecht, H., Handbuch do- gesammten Thon-
iren Industrie, Friedrich Vieweg und Sohn, Braunschweig, 1907, p. 1320.

I

JOURNAL OF THE

SiOj 50 <»)

AI2O3

PdOi 2 30

CaO 1.22

CaSO< [ .41

MgO 0.28

KjO 9.28

Na»0 8.05

H3PO4 0.0s

CO2 0.22

HaO 3.27

»

too. 13

These minerals are found in this country in Texas, South Dakota
Wyoming, Missouri, Minnesota, Utah, Arizona, and Colorado.
Albany slip is also sometimes used as a flux in chemical stone
ware.

There are many different uses for chemical stoneware an
hence the body must be varied to meet the different require
ments. Corundum2 is sometimes added to the body in cas
unusual resistance to temperature changes is necessary. I:
case the ware must be unusually resistant to chemical action-
a high degree of vitrification is desirable. For certain purpose;
the ware should be very resistant to shocks and strains and then
fore the body should be unusually tough and strong when burnec
The physical properties of the ware are as important as the chen
ical properties.

A salt or slip glaze is commonly used in the manufacture, a
though a glaze corresponding to the formula of cone 1 is som
times employed for ware which is burned to cone 4. Carborui
dum,3 because of its resistance to chemical action, is sometim
applied as a coating on chemical stoneware — water glass or boi
acid being used as a binder.

1 Clarke, E. W., "Analyses of Rocks," U. S. Geol. Surv. Bull., 591, 3

( 914)-

2 D. R. P. Xr. 158,336.

3 Kerl, B., Cramer, E., and Hecht, H., Op. cit., p. 961.

AMERICAN CERAMIC SOCIETY.

269

The Occurrence of the Clays Investigated.1
Tionesta Clay, Ellis, Ohio.
A sample of Tionesta clay, from the drift mine of Joseph Moody,
>cated one mile south of Ellis, Muskingum County, is designa-
sd as "A" in the text. This clay is shipped to the potteries at
anesville and Cambridge, Ohio, and is used in the manufacture
f art ware, mosaic tile, and cooking ware. The sample was
aken from a stock pile of the clay. The vertical section of the
line is as follows:

Ft. In.

Shale 15 o

Limestone, Putnam Hill 4 9

Clay and shale 12 o

Sample "A" Clay, light, plastic, Tionesta. ... 4 10

Lower Kittanning Clay, Roseville, Ohio.

The sample was taken from an open cut mine of the Hydraulic
'ress Brick Co., at Roseville, Muskingum County, and is desig-
ated as "B" in the text. This bed has a sufficient surface cover-
ig so as to be unaffected by weathering. A vertical section is
s follows:

Ft. In.

Shale 12 o

Shale, dark o 2

Coal, rough o 6 j

Shale o 3 I ......

~ . Middle

Coal 1 7 >

Kittanning
Clay o 72 j

Coal 1 9 j

Clay, siliceous 1 2

Sandstone, irregular bedded 3 2

Shale, gray 7 o

Sandstone, shaly 3 6

Shale, gray 6 4

Coal, shaly o 1

Clay, plastic, dark gray 5 o ]

Clay, plastic, with large nodules of

ferruginous limestone 2 of Oak Hill

Clay, plastic light 1 o |

Clay, flint, light gray 3 o J

Coal, smutty, lower Kittanning. . .0 1

ample "B" Clay, plastic, light to dark gray.. 6 S

1 The Ohio clays were sampled by Mr. Wilbur Stout of the Ohio Geol.

urvey.

2 7<)

JOURNAL OF THE

Lower Mercer Clay, White Cottage, Ohio.

The sample of Mercer Clay, designated "C," was taken ;it the
open cut mine of W. P. Oerhart, located about one mile west of
White Cottage, Muskingum County. This day is used exten-
sively by the Muskingum Pottery Co., White Cottage, and by
the A. E. Hull Pottery Co., Crooksville. Although the bed is
under shallow covering, the clay is but little weathered. A ver-
tical section is as follows:

I't In

Limestone, Lower Mercer i i

Shale o 3

Coal, bony o 2) Middle

Coal, good f) 3 J Mercer

Clay, light, siliceous 2 10

Coal, shaly o 3

Clay, siliceous, light, lower part

ferruginous 5 o

Clay, dark, flinty, siliceous o 1

Clay, plastic, light gray 1

Sample "C" I Clay, light, siliceous 1 9 I Lower

) Clay, dark, somewhat shaly 1 8 [ Mercer

Clay, light, siliceous, hard 3

Clay, light, plastic 1

Mogadore, Ohio.

The mines are located at Mogadore, Summit County, on the
W. & L. E. R- R- The sample, designated "D," was taken from
a large pile of unweathered clay representing 15 rooms in the
mine. A vertical section is as follows:

Shale, dark

Coal, smutty

Sample "D" Clay, siliceous, gray

Shale, dark, argillaceous.

Ft.
6

Lower Kittanning Clay, Toronto, Ohio.

The lower Kittanning clay, designated "E," is from the mine
of the Toronto Fire Clay Co., Toronto, Jefferson County, and
was sampled under covering. The sample is a representative
one from different rooms in the mine. A vertical section of the
mine is as follows:

AMERICAN CERAMIC SOCIETY. 271

Ft. In.

Shale 1 .5 o

Coal, Lower Kittanning 2 6

Sample "E" Clay (lower part siliceous) 7 o

Clay, very siliceous, hard, gray ... 5 o

Clay, bluish, sandy 2 o

Lower Kittanning Clays, New Brighton, Pa.

The clay, designated "F," was taken from a large pile of the
,,ower Kittanning clay at the Sherwood Bros. Pottery, New
Brighton, Pa. This sample came from under cover and the sec-
ion is similar to that at the mine of the A. F. Smith Co., both
ieing Lower Kittanning clays occurring directly under the Lower
kittanning coal. The clay at the A. F. Smith Co. plant at New
Brighton, is exposed and therefore is more weathered than that
t the Sherwood Bros. Pottery. A vertical section of both mines
5 as follows:

Ft. In.

Coal, Lower Kittanning o 1

Samples "F" and "G" Gray, plastic 5 o

Tionesta Clay, Crooksville, Ohio.

The sample designated "H," was taken from the mine of W.
L Brown and S. H. Sharky, located one mile west of Crooks-
ille, Perry County. This clay is mined by drifting and is used
xtensively in both white ware and stoneware potteries. The
ample was taken from well under cover and from a clean face.
Ahe clay from this mine is shipped principally to the A. E. Hull
'ottery. A vertical section is as follows:

Ft. In.

Shale. 5 o

Limestone, Putman Hill 1 8

Coal, Brookville o 4

Clay, light, plastic 6 o

Clay, dark, carbonaceous, horizon

of Tionesta o 2

Sample "H" Clay, light, plastic 6 o

Lower Kittanning Clay, Firebrick, Ohio.

The Lower Kittanning clay, designated "I," is used extensively
i the manufacture of fire brick and sewer pipe in the southern

272 JOURNAL OF THE

pari <>f Lawrence County. The sample was taken from an opi
cut mine at Firebrick, Lawrence County. The bed is unusual
free from nodules or sulphur compounds. A vertical section i
as follows :

IM. In.

Sandstone 8 o

Clay, sluily, Oak Hill 2 o

Coal 1 8 | Low < r

Clay 0 i1/* j Kittan-

Coal o 8 j ning

Clay, dark, carbonaceous o 4

Sample "I" Clay, light, plastic .5 o

Clay, siliceous, light 2 o

Lower Kittanning Clay, Nelsonville, Ohio.

The Lower Kittanning clay, designated "J," is used extensive!
in the manufacture of brick at Nelsonville, Hocking Count]
and the sample was taken from the mine of the Hocking Valle
Fire Brick Co., where it is used in the production of salt glaze
building brick. A vertical section of the mine is as follows:

Ft. In.

Sandstone x o

Shale 6 o

Coal, Lower Kittanning 1 8

Clay, dark, carbonaceous o 4

vSample "J" Clay, light, plastic 8 6

Sandstone, shaly 5 o

Semi-flint Clay, Scioto Furnace, Ohio.

The semi -flint clay, designated "K," was taken from a larg
pile at the plant of the Buckeye Fire Brick and Clay Co., Sciot
Furnace, Scioto County. The clay is used extensively in th
manufacture of high grade fire brick. A vertical section of tt
mine is as follows:

Ft. In.

Shale 3 o

Coal, Anthony o 1

Clay, flint 2 o

Sample "K" Clay, semi-flint, Sciotoville 1 6

Clay, "pink eye" 2 o

AMERICAN CERAMIC SOCIETY.

2 73

INVESTIGATION.

For the purposes of the investigation test pieces were molded
m seven bodies which were prepared from each of the stone-
re clays as follows:

iy No. 1 — Run of mine clay ground dry to pass a 20-mesh sieve.

iy No. 2 — Clay washed through a 150-mesh sieve.

Iy No. 3 — Residues from the washing were ground to pass a 150-mesh

sieve and then added to the washed clay.
iy No. 4 — 95 per cent, washed clay (150-mesh) and 5 per cent, feldspar.
iy No. 5 — 90 per cent, washed clay (150-mesh) and 10 per cent, feldspar.
iy No. 6 — 95 per cent, washed clay (150-mesh) and 10 per cent, feldspar +

0.16 per cent. CaC03.
iy No. 7—95 per cent, washed clay (150-mesh) and 5 per cent, feldspar +

6.1 per cent. CaC03. x

Che drying shrinkage, water of plasticity, shrinkage water,
e water, rate of slaking in water, and the transverse strengths
the green condition were determined for each body.
the softening points of the clays were compared to those of
ton pyrometric cones. The apparent porosities and volume

Fig. i. — day bars before and after warpage test.
1 Kirkpatrick, F. A., Trans. Am. Ceram. Soc, 18, 595 (1916).

(The

calcite-orthoclase eutectics were used as a guide in adding the calcium
x>nate.)

274

JOURNAL OF THE

'shrinkages of the briquettes were- determined after firing to tl
different temperatures.

Owing to the fact that commercial chemical stoneware piec<
are usually large, any tendency of the clays to warp, crack, <
break in burning may cause high kiln losses. A modificatic
of the warpage test as described by Worcester1 was thereto
employed — consisting in applying a load on the center of tl

Fig. 2. — Method of measuring de-
formation with Ames dial.

test bars while resting upon two knife edges placed 6 inches apai
Fig. i. A three-pound weight was applied to the bars having j
i" X i" cross section, corrections being made for those haviii
a smaller cross section. The deformation (Fig. 2) was express
in per cent, warpage in terms of the length of the span.

1 Worcester, W. G., Trans. Am. Ceram. Soc, 12, 818-853 (1910).

AMERICAN CERAMIC SOCIETY. 275

Transverse strength tests of the burned test bars were made
fter heating five times to 600 ° C and quenching in water at
8° C after each heating. The percentage decrease in strength
ives a rough indication of the resistance of the fired bodies to
jmperature changes.

Results of the Tests.
Sample "A," Tionesta Clay, Ellis, Ohio.

This clay slakes down rapidly in water and is comparatively
ean — leaving only 2 .5 per cent, residue on the 150-mesh screen.
: develops good mechanical strength in the raw and burned con-
ition.

As all of the bodies investigated developed greater strength
hen burned to cone 8 than when burned to cones 6 or 10, it
ould appear that, for this class of clays and bodies, the maxi-
ium strength is developed by burning to cone 8. The differ-
ice in strength is probably due to underburning at cone 6 and
/erburning at cone 10. The strength of the clays and bodies

considerably increased by screening through the 150-mesh
eve.

The vitrification range is increased by screening through the
jo-mesh sieve as compared with the run of mine clay screened
trough a 20-mesh sieve. The vitrification also starts about
vo cones lower. This is also true of the body in which the
:sidue on the screen was ground to pass a 1 50-mesh sieve and then
Ided to the washed clay. Screening the clay through a 150-
lesh sieve also increases the burning shrinkage. Bodies Nos.

5 and 6 have porosity and shrinkage curves which closely re-
;mble that of the washed clay. Body No. 7 is comparatively
sen burning and has a short vitrification range. This is eharac-
xistic of high calcium bodies. The softening temperature is
me 26+ which roughly meets the requirements of a No. 3 fire
ay as classified by Bleininger.1

The washed clay warps considerably under load. The run of
line clay, screened through the 20-mesh sieve, appears to be

1 Bleininger, A. V., "Some Aspects of the Testing of Refractories,"
roc. Eng. Soc. Western Penn., 32, 613 (1916).

276 JOURNAL <>l; THE

more resistant to the quenching treatments than the same claj
washed through a [50-mesh sieve.

Sample "B," Lower Kittanning Clay, Roseville, Ohio.

This clay slakes readily in water. Washing through a 150
mesh sieve affords a much "tighter" body than the run of min<
clay screened through a 20-mesh sieve— as shown by a lower mini
mum porosity in the first case.

The softening point of this clay is cone 25-h and does not meel
the requirements for a No. 1 fire clay. The clay warps consid-
erably under load and therefore may be unsatisfactory if used
alone.

Sample "C," Lower Mercer Clay, White Cottage, Ohio.

The clay does not slake readily in water and was therefore
wet ground before blunging. The per cent, residue on a 150-
mesh sieve is high. The strength of the raw clay is low while
that of the burned clay is good. Adding the reground residue
to the washed clay tends to open the body. The softening point,
cone 26, would probably classify this material as a No. 2 fire clay.
The warpage under load at cone 4 is excessive and high kiln
losses may be expected if this clay is used alone in the manufac-
ture of chemical stoneware.

Sample "D," Mogadore, Ohio.

This clay does not slake readily in water and was therefore
wet ground before blunging. The per cent, residue on a 150-
mesh sieve is high. The strength in the dry condition is low, al-
though the burned strength is fair. The softening point is cone
18. The resistance of the test pieces, molded from the washed
clay, to warpage under load, is not satisfactory.

Sample "E," Lower Kittanning Clay, Toronto, Ohio

This clay does not slake in water and was therefore wet ground
before blunging. The modulus of rupture of the test pieces from
the washed clay is excellent while that of the clay when burned
to cone 8 is also very good. The softening point is cone 26—
roughly meeting the requirements for a No. 3 fire clay. Th(
resistance to warpage under load is not very satisfactory.

AMERICAN CERAMIC SOCIETY. 277

Sample "F," Lower Kittanning Clay, New Brighton, Pa.

Phis clay slakes readily in water. The dry strength is low
lough the burned strength is fair. The softening point is
le 28 — roughly meeting the requirements for a No. 2 fire clay.
e resistance of the test pieces molded from the washed clay
vvarpage under load is poor.

Sample "G," Lower Kittanning Clay, New Brighton, Pa.

Phis clay slakes readily in water. The strength in the raw
I burner1 condition is fair. The addition of the reground resi-
I to the washed clay produces a more open body than the use
the washed clay alone. The softening point is cone 28, meet-
the requirements for a No. 2 fire clay. The resistance to
rpage under load at cone 4 is not satisfactory.

Sample "H," Tionesta Clay, Crooksville, Ohio.

Phis clay slakes readily in water. The dry strength is fair
ile the burned strength is good. The softening point is cone
which roughly meets the requirements of a No. 2 fire clay,
e resistance to warpage under load is not very good.

Sample "I," Lower Kittanning Clay, Firebrick, Ohio.

Phis clay slakes readily in water. The strength in the dry and
-ned condition is excellent. The addition of feldspar to this
y is advisable as it widens and lowers the vitrification range by
:ones. The color, when burned, is light gray and similar to
it of the ball clays used in the manufacture of white ware. The
tening point, cone 31, is sufficiently high to warrant the use
the clay in the manufacture of glass pots. The clay has only
mall tendency to warp.

Sample "J," Lower Kittanning Clay, Nelsonville, Ohio.

rhis clay slakes readily in water. The per cent, residue remain-
; on the 150-mesh sieve is low. The strength in the dry and
rned condition is excellent. The softening point, cone 30,
jssifies the material as a No. 2 fire clay. The resistance to
'rpage under load is fair.

JOURNAL OF THE

Sample "K," Lower Kittanning Clay, Scioto Furnace, Ohio.

This clay does not slake readily in water and was therefore \\»
ground in a bail mill previous to blunging. The dry strengj
is low, although tlu- burned strength is excellent. The softena
point, cone 32 + , classifies the material as a No. 1 fire clay. Th
clay is considered as being one of the typical fire-brick clays <
Ohio. The resistance to cracking and warpage under load i
cone 2 is very poor.

Summary.

With but few exceptions, the screening of the clays through
150-mesh sieve improves the strength in the raw and burned coi
dition, increases the dry porosity, lowers the vitrification ten
perature, widens the vitrification range, decreases the minimui
burned porosity, increases the burning shrinkage, and increas«
the density in the burned condition.

When ground and screened to pass through a 150-mesh siev
the residue from the washing may be added to the washed cla
to advantage in most cases. The addition of feldspar alone (
with a small percentage of calcium carbonate lowers the vitrific;
tion temperature. When a large percentage (6 . 1 per cent.) <
calcium carbonate is added, the bodies are more porous at tl
lower temperatures, the vitrification temperature is raised, an
in general the vitrification range is narrowed.

The maximum strength of all of the clays tested is develop*
by burning to cone 8.

There appears to be no relation between the softening temper
ture and the resistance to warpage under load at the low ter
peratures — the most refractory clays sometimes fail before t
less refractory ones.

There is a large decrease in the mechanical strength when t
clays are being burned above cone 1. In one case, the modul
of rupture was 270 in the dry state and, although the clay is
No. 1 fire clay, its modulus of rupture at cone 4 was decreas
to 27.

The run of mine clays, screened dry through a 20-mesh sie
appear to be more resistant to the quenching treatments th
the same clays when screened through a 150-mesh sieve.

AMERICAN CERAMIC SOCIETY.

279

Table i. — Miscellaneous Properties.

u

a
a °

25

6

9

12

7

9

12

7
1

3

o

27 .0
29.6

23.8
28.1
23.6
257
23.2
28.5
19 4
27 .2

23 4
29.7

24 -5
27.9

23. 1
29.0
24.8
256
20.8
29.2

18.5
28.4

83

IS -~

s

389 .5
467.9

218.6

3'95

132.0

2507

1425
259.0

384.3
532.1
194.2
191 .0

194 5
321 .0

178.8

3I4.9

3249
499.2

247 -3

350.I

93-8

270.0

Per cent, warpage.

— ri -t

U 0 V

a c a
000

o o o

83 2.42 8

77 513 9

75 6.10 15

80 3 . 48 6

48 2.75 6

58 5.91 10

58 3.22 6

30 3 89 7

83 0.84 3

58 ... 5

m ° £

a a «
w <u E

a 0

__ O <-.

'-

,3»<2

.sS 5

58 8 . 85 broke

9i 5
963

775
97-7

77-6
93 o

82.0

93 o
78.6

92 .0

82.5
96.0

835

94 o

73-5

93 5
855
96.5

775

95 o

55-5
95-4

a a
a o
SO

26 +

26 +

26

18

26—

28

28

28

31

30

32 +

-Per Cent. Apparent Porosities of Clays and Mixtures
Burned to Different Temperatures.

o. Cone 1.

& 1938

* I?-

\ 19.

\ 19.

\ 19

\ 18.

\ 24.

Cone 2.
15 90
15 88
16. IO
I450
14 65
14. 10
19.80

Cone 4.
3 -70
O. 14
o. 14

0.54
o. 14
0.03

6 49

Cone 6. Cone 8. Cone 10. Cone 12. Cone 14.

O.78
O.28
0.2I
O.I2
O.I7
O.06
I .07

0.93
O.36
O.36
O.OI
O.08
0.15
5-99

0.20
023
o. 16
0.06
0.09
0.00

2 .20
4.00
1 .50

3 90
035
o. 16
5 70

J So

Ji >i RNAL OF THE

\(, Cone 1

IB is 93

_'B 2< i 60

3B 21.45

(H ... 16.75

5B i7 5<>

6B 17.15

7^ 23.15

1 C 1 9 . 60

2 C 19.90

3C 24.00

4C 21 .55

5C 22 .OO

6C 2 2 . 05

7C 28.63

iD 24.75

2D 22.55

3D 27.48

4D 23.15

5D 24.35

6D 22.85

7D 29.05

iE 22 .20

2E 24.30

3E 24.70

4E 24.90

5E 21 .20

6E 24.80

7E 23.65

iF 20.20

2F 24.95

3F 25.60

4F 21 .90

5F 23.55

6F 20.75

7F 27.35

iG 21.95

2G 21.35

3G 23.00

4G 2 1 . 00

5G 22.30

6G 24.00

7G 27.00

Tabi 1

1 ( 'onHnued).

1 one

i I.

( '.

Cone 10.

'7 15

[2 .99

12 7"

1 2 48

1 | ,-;. .

C2 Bo

5 80

1 4 1 5

O.9I

0.24

4 97

3 20

6.30

[4.60

0.24

O.94

3 99

1 04

6.80

i.s 50

0 34

O.I5

2 03

O.78

4 ,60

4 3.5

1 4 65

O . (2

(i [2

4 77

0.60

6 40

14.28

O 99

0 J 4

2.21

2.18

5 00

23. 15

1 5 90

0 OO

0.28

0.00

5-5"

22 .80

14 55

IO.40

7-32

4-95

5. '"

« si

18.50

6-59

2-75

0.78

1 45

6. 10

8 31

24. 10

12.65

9.00

1 31

4.41

6 70

9 6fl

21-35

10.25

4 30

1 93

2-34

3 6< >

872

18.35

9.70

4.60

2 .09

1 54

6.00

6 50

21 25

IO-33

3-75

0-59

2 94

6.00

7 40

29.20

18.10

8.00

0.22

0.00

7 00

25-25

18.80

I5-30

12.12

7-33

4.80

5 H

23.60

17-50

9.40

3-63

3-37

3 37

1 1 .70

26.30

19.00

10.60

2 .92

2-75

6.20

10.48

23 -55

16 20

4 30

2 93

2.17

8.60

10.20

22 .30

7-73

7 50

2.49

2 .29

9.80

6.12

22.85

9-33

8.88

2.44

1 55

9.00

7 37

30 . 30

14 23

ii-53

9-35

9 30

19.00

19-75

io.74

10.88

10.93

10.51

1 1 .00

10.70

20.25

7-43

4.00

3-25

7.80

8.80

8.72

1983

6.50

3 00

2.21

6.95

8.50

10.40

20.55

8.60

2.50

0.23

11 .70

6.20

17.60

5 -98

1 .no

5 69

5-82

9.20

21 .00

1 1 . 69

4 50

2 .46

12 . 10

6.30

2-41

2315

1425

5-75

0.06

0.00

7.90

8.20

23.40

ib. 58

14 25

1324

12-45

7 50

2 44

25 05

1 1 .00

3-75

1 .03

5 57

6.40

6.68

27 .00

I3-70

5-4°

3 3i

3-71

5.10

5 58

24 55

7.89

4.00

2.58

6.88

10.30

3.00

25.20

6.36

4.00

4 44

4.88

9.80

2-77

24.30

8-33

2.50

0.29

1 .84

3 -50

2 .90

31-65

14 35

4.00

0.08

0.28

1 .70

22.95

20.00

15.00

10.32

9.60

5-20

10.85

1985

16.00

5.00

1 .80

7-49

8.00

11 .20

21-95

18.00

8-75

3.02

5.60

7.10

12 .60

18-75

14.00

2.80

0-39

0.26

4. 10

10.90

18.95

14.00

4.00

115

5-23

6.60

5 09

21 .20

12.50

5.00

0-73

6-55

8.10

11 45

26.70

15-95

4.00

0-33

0.18

9.40

AMERICAN CERAMIC SOCIETY.

2S1

No. Cone 1.

iH 24.95

2H 26.75

3H 25.OO

4H 25.65

5H 22.75

6H 24.00

7H 27.00

il i9-i5

2I 18.23

3l 1965

4I 18.30

5l I4-48

61 16.20

7I 16.70

ij 21.75

2 J 24.20

3J 24.95

4J 22.80

5J 22.85

6J 22 .90

7J 27.00

iK 23.85

2K 21 .65

3K 25.30

4K 22 .90

5K 21 .70

6K 20.55

7K 16.35

T

\BLE 2.-

—{Conti

nued) .

Cone 2.

Cone 4.

Cone 6.

Cone 8.

Cone 10.

Cone 12.

Cone 14

2295

17.80

16.OO

14 SO

1 1 .40

IO.OO

5-48

22 OO

16

20

5-50

025

330

0.59

7-95

19. IO

14

58

10.60

6.68

414

i-79

2.42

21 . 20

7

85

2 .OO

037

O.OO

4.20

18.OO

7

98

i-75

0.41

0.17

5 -30

7 56

21 .20

10

23

2.50

o.35

O.29

059

631

26.7O

15

95

4 50

0.29

O.24

0.78

19. IO

16.23

12 .00

10.78

7.84

6.30

6.64

I9.4O

11 95

7-25

523

4-55

0.17

8.50

19. IO

12.55

7-75

5 54

4-74

1 30

8.90

17-55

9.29

3-75

0.61

o.45

1.80

5 04

I340

589

1 .90

0.08

O.OI

2 .70

3-91

I5-30

10. 10

4-75

o.33

115

o.49

4-52

14 65

1327

8.50

3-22

0. 12

0.00

9. 10

18.95

15.01

12 . 10

IO.66

10.45

8.30

2 1-75

12 .40

3.00

O.08

o.43

1 .06

5-77

21 -55

1 2 . 30

4-75

O.23

0.32

360

8.71

I8.20

IO.7O

2 .60

O.O9

0. 19

o.73

7-43

I7.3O

9.80

2.25

O.O9

0. 12

3 30

7.85

18.7O

9.60

2.25

O.O9

0.13

3-31

7.40

2715

1745

7 .00

O. 12

005

0.02

10. 70

20.85

22.50

22 .80

21.65

19. IO

17 .20

18.15

18.55

I4.O7

9-75

8.70

8.40

2.30

2.85

I9-65

18.73

12.50

11.25

10.70

5-25

3-57

16.15

I2.98

9.00

6.65

5.67

1 .60

342

HoO

11.66

7 50

2-95

1.49

4.72

1 .90

I53O

9-53

8.00

6.47

5.38

0.00

1. 81

17-^5

15

65

12 .00

IO 14.

8.82

0.15

1 S3

Table 3/ — -Modulus of Rupture of Burned Bodies.

No. Cone 6. Cone 8. Cone 10.

I A 4624.0 6680.0 3812.0

2 A 7200.0 8265.0 5330.0

3 A 7240.O IO747.O 4671.O

4A 6538.O 7780.O 4910.O

5 A 6540.0 6787.0 5966.0

<>A 6616.0 8728.6 6001.0

7A 5566.0 6958.3 57790

282

JOURNAL OF THE

Nil

iB..

jB..

jB

4B

5B..

6B.

7B..

iC.

2C.

3C.
4C.
5C.

6C.
7C.
iD..

2D..

3D..
4D..

5D..

6D. .

7D..

iE..

2E..

3E..

4E.

5E..

6E..

7E...
iF..
2F..
3F-.
4F..
5F..
6F..
?F..
iG..
2G..
3G..
4G..
5G..
6G..
7G..

\ltl.l

1 Continued).

Com 6,

t one 8.

Cone 10

3I90.0

3416.2

J()(>S 0

8556 0

8680 0

7504.8

76280

95050

9462 ,0

813O.O

9738.3

691 I (>

7010 <i

8482.9

7050.0

6957.O

75140

6494 0

5434 0

6003 . 7

4853-0

2964.0

3223.7

[670.0

57030

7285.0

4681 .0

4534-0

6485.0

4543 0

52540

5843 7

4421 .0

.5491 0

7802.5

4749 0

4564 . 0

5090 0

4321 .0

4090 . 0

7067 . 1

3464-0

2023 .0

2283.7

2181 .0

4666 . 0

4820.0

3361.0

3686 . 0

4585*0

2418.0

4284.0

5581.4

3326.0

4621 .0

5790.0

2991 .0

4606 . 2

5877 -5

3944 0

3944 ■ 3

4627 -5

2218.0

2788.7

3568.7

2260.0

7904 -3

6847.5

5690.0

5882.8

6397 5

5084.0

5128.6

5191.6

4584.0

5430 0

5588.5

4I55-0

5168.7

5385 -o

48730

291 1 .2

3063 0

1 909 . 0

4932 5

58150

3368.0

4213.8

5133 0

3750.0

44725

5507-0

3869.0

6275.0

6727 .0

2675.0

4751-4

6059.0

3260.0

39o8 . 7

5309 0

2057.0

2580.0

3288.7

1521-0

5ii3 0

5183.7

4474.0

4948 ■ 7

6350.0

2643.0

4993 • 8

5473.0

5400.0

4853 ■ 7

6284.0

4506.0

4973 0

578o.o

4381.0

37150

58900

4545 0

AMERICAN CERAMIC SOCIETY.

283

No.

iH

2H

3H

4H

5H

6H

7H

il

2I

3l

4l

5l

61

7l

J

2j

3J- ....

4J

5J

6J

7J

iK

>K

3K

4^

5K

6K

7K

BLE 3-—I

Continued) .

Cone 6.

Cone 8.

1974.O

3043 • 7

66 1 5 . 0

7148.7

2873.3

6837 .0

5682.0

7960.0

6 1 1 1 .2

9075 0

6591 .2

7768.5

5058 . 7

6208.7

4068 . 7

5075-0

10448.3

7190.0

9438 0

7045-0

9691 .0

67340

7227.5

6847 . 1

8837 . I

S56I.O

6052.9

1569.0

3507 ■ 5

4699 . O

6150.0

5619.0

6522.5

55230

8640 . 0

6404 . 0

6635.0

6165.0

8481.2

5561.4

6221 . 1

1649.4

2129.3

3655°

7411 .0

3730.0

8976.0

2836.6

5742.o

3585-0

4710.0

2235.0

8740.0

2437-8

5621 .2

Cone 10.
1956.0
3689.O
1912 .O
4586.0
3941 o
5535-0
3682.O

3014.5

6i75 6
6073.0
7023.0
7207 .0
7238.0
5169.0

2034.0
5004.0
4125.0
4018 .0
5799-0
5850.0
4528.0

2041 .0
2716.0
3201 .0
3546 .0
35670
2046.0
5064 . o

able 4. — Per Cent. Volume Shrinkage of Clays and Mixtures Burned
to Different Temperatures.

No. Cone 1. Cone 2. Cone 4. Cone 6. Cone 8. Cone 10. Cone 12.

1 A 13.20 15.90 24.00 24.31 23.30 24.90

2A 19.80 2I.IO 29.55 29.50 28.80 29.10 21.80

3A 20.90 22.05 30.85 3225 32.35 32.55 28.IO

4A 17 95 20.83 28.09 3025 35.60 28.00

5A 17-53 19 35 27.41 28.75 28.85 28.85 25.80

6A 21.15 23.80 30.64 31.60 31.45 31.55 28.50

7A 1385 14.85 24.44 23.51 17.20 ... 22.90

284

JOURNAL OF THE

T. mi 1.1:

No. I one 1 I on<

1 H 14 411 147"

2H 20.45 25 .05

3B 21.85 25.6.5

4B 19 75 22 .60

5B 15- i« 20.65

6B 20.00 22 .45

7B 14-5° 15.80

iC 743 8.36

2C 16.15 20.25

3C 10.08 13.89

4C 12 .60 14.87

5C 12 .83 18.00

6C 12 . 23 15.12

7C 6.10

iD 3.70

2D 8.43

3D 586

4D 8.18

5D... 8.55

6D 8.33

7D 2.93

iE 714

2E 10.64

3E 10.85

4E n.55

5E 1440

6E 1045

7E 10.02 •

iF 3.64 5.92

2F 9.23 12 .85

3F 8.10 11 . 36

4F 1 1 .28 14.03

5F 9 23 13 .96

6F 9.78 13. 17

7F 593 8.28

iG 6.48 737

2G 9.49 16.74

3G 8.53 9

4G 9 73 14

5G 11.68 15

6G 10.78 14

7G 5 59 8.05

4. — (Continued).

< one t. Cone (,.

66

98
64
5i
36
98
82

54
60

87
21

65
47
19
56

17.12

32.30
32.40
29 00

27- 13
29.18
19.24

I.5-I4
26.75
20.88
22 . 10

16.15

6.92

17-85

18.13
17.48
I9-38
16.53
15-03

1503
23-50
24.80
22.85
24-55
21.88
14.28
825
19.00
16.25
20.00

12 .00
14.00
20.20
18.10
20.55
24-30
21 .70
1765

[6.53

31 i';
34 -58
28.73
27.48
28.80
21 .70
16.25

27-95
26.65

2425

18.50 25.36
20.70 24.25

22 .50

7-95
96

9-43
22 .70

9.42
22 .00
20.00
25.69

20.00 26.38
13.25 24.65
20.03
14.40
24.00
24.30
24.00
26 . 00
26.60
22.75

Cone 8.

<i ah

24.80

•6 60
26.80
22 .40

24 15
24.80

1 6 . 40
28.40
28.40
24.85
23-55
27-35
25-75
9.00

23-45
24.90
23.40

22 .4

23 65
21 .97

6.06
21 .90
24.80

25-55
20.35
23-35
23.00

12.43
28.65
28.25
24.40
22. 15
28.15
28.45

Com 10.

X . 63

M, (o
26.80
23.80
24-65
23- IO
23.4O
25.9O
IO.9O
23-40
25.00
23.4O
22 .30
23.60
20.03

5 57
20.60
22 .40
23.00
18.30

23-50

H-95
22.93

27-75
19.20
20.25
27.80
29-40
20.20
23.00

23-45
22.95
22 .70
22 .00

25 -55

Cone 12.
6.3O

4'-25
?,?> • 95
22 .20
27.20
31 -6o
24.60

13-70
19-35
20.60

23-75
17 30
19 50
I4-30
14. 10
19.80
20.50
I4-50
5 40
9.00
19.00
4.60
I5-90
17 .60
20.30
14. 10
16.75
IJ-30

7-50
3 40
5.10
10.30
9.80
3 -SO
1.70

18.00
20.60
23.20
1633
15-45
18-45
12 .40

AMERICAN CERAMIC SOCIETY.

285

Table 4. — (Continued).

Cone 1.

Cone 2.

Cone 4.

Cone 6.

Cone 8.

Cone 10.

Cone 12

6.63

8.32

H-35

13.OO

I370

13-43

13 . IO

1388

16.52

23-55

23.OO

28.55

33

.40

I5.00

16.68

29.70

32 .OO

32.85

33

.00

34.OO

12 .70

16.36

25-35

26.40

27 . IO

27

■30

21 .OO

13-45

18.54

25.00

27.30

26.80

27

.00

I9.OO

13-55

16.87

23 .00

27 .OO

28.80

28

85

28.60

9.60

II.44

18.15

25.OO

28.20

28

20

24.80

I3.63

I503

16.80

19.60

2O.4O

20.65

22 .OO

17-75

19.40

22 .30

25 OO

25-75

2595

26.4O

17-53

17.71

21 55

22 .OO

22 .OO

22 .70

27.80

I4.30

16.38

21 .OO

24.OO

2425

22 . 80

19.17

21 .40

24.2O

25-70

26.25

28.60

I5.OO

19. IO

20.98

24.05

27 .OO

28.15

27.70

9.60

II.44

l8.I5

22.50

24.60

26.90

22 .90

978

IO.96

I330

12 . IO

IO. IO

10. 10

7-50

13-73

16.4O

22.45

29.OO

30. 10

29.90

28.9O

I4.9O

16.9I

22 .80

28.50

29.70

29.00

23-70

I3.60

18.23

23-55

29.OO

30.45

3040

28.OO

I550

18.55

24.00

28.IO

29.25

29.50

2I.IO

15 90

I7.84

25.15

29-50

30.30

31.90

25.60

7-32

8-45

1450

22 .OO

26. IO

26.30

25-50

I5.40

16.59

18.40

20.25

20.90

18.88

17 . IO

23 05

2515

28.90

31.25

32 .00

32

30

33 40

21 .40

23.28

24.90

27.50

30 . 45

30

85

33-55

22.75

27.63

30. IO

32 .OO

3355

34-

15

35-6o

23.20

28.78

30. IO

32 .OO

34- 10

34-

75

35-30

20.80

26.18

28.20

30 . 50

32.15

32.

60

34.00

23 ■ 70

23.80

26.OO

28.30

28.

50

30.15

Table

-Shrinkage Relations.

u 3

2 V

2 «E

Ji.0

a Oh

Run of mime 78.8

Washed 83 . 9

Run of mine 7 1 . 3

Washed 90 . 6

.seg

.c «o
■" >

-.9 >>

» a v

« 2
&*"

41-7
44 o
40 3
48.5

O —
CO

SfcE

8S3

a.

37-3
39-9
31 .0
42.1

Sum

h* s*

OJfl

■Srt'S

rt > <n

A

0.92

0.91

0.77
0.87

.5 5

u **
u

a,

52.9

52.4
565
53 -5

- — :
30.3

34-2

27.5
34-8

SB

cB

8.3 .»

30.5
33-2
30.8
33 -0

AMERICAN CERAMIC SOCIETY.

iC

2C

iD
2D
iE

2E

iF
2F
iG
2G
iH
2H
il

2l

ij
iK

2K

Run of mine

Washed ....
Run of mine.

Washed ....
Rune of mine

Washed... .
Run of mine

Washed. . . .
Run of mine .

Washed. . . .
Run of mine .

Washed. . . .
Run of mine .

Washed. . . .
Run of mine.

Washed. . . .
Run of mine .

Washed. . . .

iii.i-. 5

— (Continued

i- II

V—

v- >,

u ■

0**rr

•— if.

u '

s =

\t 0

8 *

s =

a

°i

SI-

*

e £ >

A .,, -J

>

U ^

a-o

ft5.

r u ~

3 2 S

S 0 Si

4 /

1 a

X. ii
'/. *- u'
D

■ c ~

'S - .
> il

.5 a

- if

S"*JS

IILU

= E E

— - d

= is

■v -

a 0 a,

° a
c —

0 a, C

.2 a jj
« s »

0 ""a

•=-3

c

Ph

Ch

K

£

£

£

6i . i

30 2

30.9

1 .02

49-5

23 5

22 .9

8 1. 6

38

3

43

3

1. 1 3

46

9

3i 4

26 6

49 -8

20

8

29

0

1 .40

4'

7

19. 1

16. ]

75-0

37

1

37

9

1 .02

49

5

28.9

26.I

49 .o

23

5

25

5

1 .07

48

0

18.9

18.9

86.3

50

5

35

8

0.71

58

5

33-2

36 2

54-6

26

0

28

6

1. 19

47

5

21.0

20.0

83.8

38

8

45

0

1. 16

46

3

322

26.4

64 5

33

7

30

8

0.91

52

2

24.8

25 1

76.0

39

8

36

2

0.91

52

3

29.2

~77 4

631

29

8

33

3

1 . 12

47

1

24-3

22.7

88.0

46

0

42

0

0.91

52

3

33-8

323

88.5

56

0

32

5

0.58

63

3

34-2

40.6

89.6

55

6

34

0

0.61

62

0

34 6

41.2

59 6

30

6

29

0

0-95

51

3

23.0

22 .7

87.6

48

5

39

1

0.81

55

4

33-7

34-7

43-5

17

8

25

7

1.44

4i

0

16.8

14.O

86.0

48

0

38

0

0.79

55

8

33- 1

34-6

The properties of the above Ohio and Pennsylvania stonewai
clays were studied by the Bureau of Mines under the direction (
Prof. A S. Watts in cooperation with the Ohio Geological Survey
The Pennsylvania Geological Survey also assisted in obtainin
the Pennsylvania stoneware clay samples. Although this wor
was done with special reference to chemical stoneware, man
results were brought out which may be applied to a number <
other ceramic industries.

3E USE OF DENSE SOLUTIONS IN DETERMINING
THE STRUCTURE OF PORCELAINS.

By H. Spurrier, Detroit, Mich.

Petrographers and mineralogists have used dense organic and
Drganic solutions for a number of years in the separation of
nerals having different specific gravities. Under certain con-
tions fused solids have also been used.

The use of a dense solution in the grading of porcelains was
ggested by the variations in specific gravity encountered in
derfired, properly fired and overtired spark plug insulators,
slightly underfired piece of porcelain has a higher specific gravity
an a properly fired piece of the same composition while the
jcific gravity of an overtired piece is less than that of a properly
ed piece.

In an attempt to determine the reliability and accuracy of a
sual inspection of porcelains the following tests were made:
Four pieces of porcelain were selected and visually classified
a competent inspector. Two of the pieces were rejected by
n as being overtired and two were passed as being properly
2d. However, one of the pieces selected as being properly
id showed a tendency toward slight overfiring. Upon sub-
tting the same four pieces to another inspector, he classified
zm in the same way, confirming the judgment of the first in-
;ctor.

Careful specific gravity determinations on the four pieces gave
! following results:

Classification
(inspector). Sp. Gr.

Overfired i .865

1.861

Properly fired 2.086

1. 916

The piece having a specific gravity of 1.916 was the one in re-
rd to which there was a slight question as to its being properly

JOURNAL OF THE

fired. It is apparent, therefore, that visual examination by skill
ful inspectors is not all that could be desired in the selection o
the properly fired porcelains.

For the rapid laboratory or factory testing of porcelains b]
means of heavy solutions, a liquid having the following propcrtic
is necessary :

i. The specific gravity should be 2.3 or above.

2. The liquid should be mobile at the necessary concentrations

3. It should not exert a corrosive action on the material beinj
tested.

4. It should be fairly constant in specific gravity.

5. It should be available in commercial quantities at reason
able cost.

In endeavoring to find a liquid which would be suitable for th(
specific-gravity determinations on porcelains, the following ob
servations were made in reference to some of the solutions usee
by mineralogists :

Acetylene tetrabromide can be diluted only with ether or ben-
zol and is therefore somewhat unstable as to specific gravity.

Methylene iodide is easily decomposed by light.

Klein's solution, cadmium borotungstate, is prepared with con-
siderable difficulty.

The specific gravity of concentrated zinc chloride (82.12 pei
cent. ZnCl2, 17.88 HoO) is 2.2307. At the desired concentration
however, it is quite viscous.

Potassium mercuric iodide meets all of the requirements of z
liquid to be used for this purpose but is quite expensive.

Acid nitrate of mercury is quite mobile and remarkably stabl*
and appears to have all of the desirable properties enumeratec
above.

In the grading of porcelains in our laboratory by means of th<
acid mercuric nitrate, two solutions are used, the one being dilutee
so as to just float a properly fired piece and to allow an under
fired piece to sink — the other being so diluted as to allow a properb
fired piece to sink and an overtired piece to float. The solution
may become a little too dense with age but this is remedied fr
the addition of a little water acidified with nitric acid.

AMERICAN CERAMIC SOCIETY. 289

The density of the solution, when once established for a certain
)dy, should be checked from time to time by means of a good
ecilic gravity hydrometer. Upon removing and washing the
sted pieces with water, a yellow deposit will occur but this can
: easily removed with a little acetic acid.

It is, of course, quite obvious that the specific gravity of fired
are may vary quite sensibly in consequence of the presence of
echanical voids, and that a properly fired piece of ware may
►ssess an apparent specific gravity below the established stand-
ds. In such cases the ware is rejected by the solution. This
as it should be where careful selection of the porcelains is neces-
ry. The fact that invisible imperfections in the porcelains are
once detected by this method makes it more efficient than visual
amination by the inspector.

We have used a solution of the acid mercuric nitrate only as a
pid laboratory check for the insulator inspectors but its use
Fers interesting possibilities as an exact method for factory
sting.

By the employment of this method, one more operation is
lieved of the uncertain personal factor and placed upon an un-
rying mathematical basis.

Jeffery-Dewitt Co.,
Detroit, Mich.

DISCUSSION.

Mr. McDougal: I would like to ask Mr. Spurrier if he has
und that a difference in the thickness of the glaze on a spark
tig insulator would be an important factor in the employment
this test; that is, would a heavier glaze on one piece affect the
parent specific gravity so as to indicate an apparently denser
ucture than that of a piece having a lighter glaze coating?

Mr. Spurrier: A glaze usually has a higher specific gravity
an the body to which it is applied, and if the glaze area is very
ge, in proportion to the volume of the porcelain, it will
a source of error in the specific gravity determination. The
ize, if uniform, is usually so thin that, within the necessary
fits, it does not affect the behavior of the piece in the solution.

290 JOURNAL OF THE

Mr. Crbighton: This method, in its simplicity, appeals to
me very greatly. I should like t<> ask Mr. Spurrier if the size <>i
the piece is not an important factor to be considered, and also
if the character <>f the surface- of the piece, whether rough or smooth,
will not afTect the results? If the surface of tin piece is rough,
will not a certain amount of air be trapped when the piece is
dropped into the solution — thus affecting the reading? If the
piece is large the importance of this factor would be diminished.

Mr. vSpurrier: In making specific-gravity determinations in
water great care is necessary to prevent the entrapping of air bv
the piece being tested. I have found, however, that by exercising
reasonable care, no difficulties from entrapped air are encountered
in the testing of the porcelain pieces in the dense solutions.

Mr. Watts: I would like to ask Mr. vSpurrier whether
in the use of these solutions — there would be any objection to
placing the porcelains in a metal gauze basket while inserting
into the solutions? Are the solutions corrosive to metals — -nickel,
etc.?

Mr. Spurrier: Yes, the dilute nitric acid, but some of the
high silica irons now available are particularly resistant to nitric
acid and the mercury does not affect them.

Mr. Watts: That is just the point I wanted to raise: would
a small amount of the metal that might be taken up from the
basket materially affect the density of the solution?

Mr. Spurrier: No, I think not, because the volume of the
liquid is large in proportion to the volume of the pieces being
tested.

JOURNAL

OF THE

AMERICAN CERAMIC SOCIETY

A monthly journal devoted to the arts and sciences related to
the silicate industries.

Vol. 1 May, 1918 No. 5

EDITORIALS.

CHEMICAL STONEWARE.

The manufacture of chemical stoneware has been greatly in-
fluenced and stimulated by the demands of the war and the
importance of this product in the manufacture of certain war
essentials, such as acids, heavy chemicals, high explosives, medi-
cines, etc., has proven second to no other product. In the new
and important industry of the fixation of atmospheric nitrogen,
which the Government is developing upon such an enormous
scale, chemical stoneware will play an important and leading
role. Chemical stoneware is used on a large scale in the produc-
tion of picric acid and of nitric acid. Perhaps no development
of the war holds as much popular and scientific interest as the pro-
duction of poison gases and gas shell. Here again chemical
stoneware is necessary to large scale production and without
it the manufacture of this new war essential would be greatly
handicapped.

When we consider that chemical stoneware really takes the
place of chemical glassware in large scale production, its uses
and necessary properties are better understood. In the labora-
tory the chemist uses chemical glassware for his work but the use
of chemical glassware in factory operations is out of the question.
Here, chemical stoneware enters and its manufacture into the
complicated shapes required is little short of wonderful. In
order that it may stand up under factory conditions — which may
include corrosion, high temperatures, high pressures, and rapid

2Q2 JOURNAL OF THE

heating and cooling -the quality and construction of the ware
must be <>f the very highest grade.

The entire system of an acid plant is often made foni this ware.

Starting from the still, all of the pipes, cooling coils, pumps, ele
vators, containers, towers, condensers and collectors are so con-
structed that the acid is in contact only with acid-proof stone-
ware throughout the system. The automatic acid pumps have
valves and pistons of stoneware — ground to fit accurately. The
exhausters have stoneware casings, propellers and driving shafts
accurately fitted with ground joints. To complete the cycle, the
acid will soon be conveyed to the consumer in chemical stoneware
carboys.

Before the war, it was generally believed that the best stone-
ware could be made only in Germany — where the development
of the industry had received much care and attention. Some of
the highest grade stoneware on the market in this country at
that time was made from raw clay bodies imported from Ger-
many. With the outbreak of the war, however, it became neces-
sary to develop sources of raw materials and this has been done
to a highly successful degree. At the present time it may be
stated that there is being produced in this country chemical
stoneware which is superior in every respect to any made in Ger-
many previous to the war. The manufacturers of this essential
product have met the increased war demands in splendid style
and all requirements have been quickly met. Chemical stone-
ware will undoubtedly continue to play a leading role in our greatly
expanded chemical industry.

THE AMERICAN CERAMIC SOCIETY AND THE WAR.

It is gratifying to note that the American Ceramic Societ)
through the efforts of its War Service Committee, has receive
recognition by the War Industries Board through the appoint-
ment of Homer F. Staley, President of the Society, as Technical
Advisor in Ceramics to the Board. Mr. Staley will cooperate
with the other members of the staff dealing with the problems
arising in the production of such ceramic products as optical glass,
chemical stoneware, tableware, chemical porcelain, building

AMERICAN CERAMIC SOCIETY. 293

materials, and the production and importation of white-burning
:lays.

The importance of ceramic products in the manufacture of a
'reat many of the war essentials is unquestioned. Up to this
time our members have individually and collectively given their
services toward the solution of the many war problems arising
in the industry but there has been a serious lack of team-work
md cooperation. Mr. Staley's appointment as Technical Ad-
visor to the War Industries Board fills a long-felt want and gives
to the American Ceramic Society, the one great technical society
representing a large and very essential industry, the recognition
t deserves.

WAR CURRICULUM IN CERAMIC ENGINEERING

It is interesting to note that the Committee on Ceramic
Chemistry of the National Research Council is now preparing a
War Curriculum in Ceramic Engineering for those Colleges and
Universities maintaining units of the Student Army Training
Corps and having Departments of Ceramic Engineering. After
consultation with the Ceramic Departments involved, a tentative
curriculum has been agreed upon and courses in this branch of
engineering will be afforded. The Curriculum in Ceramic Engineer-
ing as outlined includes most of the courses now offered. How-
ever, special courses in Combustion and Fuel Engineering are
notable additions. It is to be hoped that a sufficient number of
the students will take advantage of the Ceramic Engineering
courses so that the future demand for technically trained Ceramic
Engineers will be supplied.

ORIGINAL PAPERS AND DISCUSSIONS.

A STUDY OF DRAFT MOVEMENTS IN FLUES.

By Edwin H. Fritz, Derry, Pa.

Introduction.

In a paper by C. B. Harrop1 on "The Inadequacy of Static
Pressure Draft Gages," he proved beyond question that the
ordinary draft gage is apt to be misleading, and that its readings
do not indicate the rate of movement of gases in the kiln because
the pressure, producing this movement, is entirely different from
the static pressure measured by the draft gage. The pressure
causing the rate of movement or velocity of flow of gases is called
the "velocity pressure" and when added to the static pressure —
which is the pressure required to overcome the resistance offered
to the flow — gives the total pressure or draft intensity of a stack.
The "velocity pressure" is not apt to vary directly v\ith the
static pressure and for this reason cannot be equal to it. There-
fore, if the static pressure reading, as obtained by a draft gage,
is assumed to be the velocity pressure, the results will be erroneous.

As an illustration, Harrop discussed a stack having a grate bar
furnace built into its base, and with the internal gases at a tem-
perature cf 600 ° F. If the fire and ash doors were closed, the draft
intensity was 0.6 inch (water column). If the ash doors were
opened and the friction in the chimney was considered as negligible,
a velocity pressure of 0.3 inch was developed. At the same time, the
draft gage would show a reading of 0.3 inch over the grates. As-
suming a constant stack temperature and a thicker fuel bed offering -
50 per cent, more resistance to the passage of the air — only three- '
fourths as much air would pass through and the corresponding
velocity pressure would be less than before. The draft gage
1 Trans. Am. Ceram. Soc, 18, 223 (1916).

AMERICAN CERAMIC SOCIETY. 295

ecording the static pressure, or the difference between the total
raft intensity and the velocity pressure, would read higher,
^hus he pointed out that by thickening the fuel bed and, there-
ore, by decreasing the volume of gas flowing through the grates,
he draft gage gave a higher reading. This higher reading would
rroneously indicate a better draft.

Investigation.

It seems that the inconsistency and tendency of the draft gages
o mislead is not quite generally known and as there are many
n actual use the necessity for a better method for its use is evi-
lent. Accordingly, this investigation was undertaken with a
riew to either finding a way to make the draft gage indicate the
ate of movement of the gases or to find some other way of measur-
ng the volume cf gas passing through a flue.

The method selected was that suggested by Harrop1 at the
:lose cf his paper and also mentioned in the discussion of his
)aper. Briefly, the measurement of the static pressure between
wo points in a flue was taken as a basis — the friction between the
:wo points always being constant with any given velocity of
rases.

The apparatus used consisted of static pressure tubes inserted
nto a flue at the two points and each connected to one end of a
J-tube, as used on the draft gage. A differential reading giving
Jie friction between the two points was then taken. As long as
10 air passes through the flue, there is no friction and consequently
Jie liquid in both tubes would stand at the same level. How-
ever, as soon as air passes through, there is friction and this friction
would be measured by the differential reading on the gage. It is
evident that the differential reading would vary directly as the
volume of gas passing through the flue, since the greater the volume
the greater the velocity of the gas and the greater, therefore, the
resistance offered by the flue. However, since the velocity of the
jas in the flue between a kiln and stack is only about 600 feet per
minute, the two points in the flue would have to be a great distance
apart in order to get an appreciable reading — a straight flue
offering but little resistance. In fact, the full length of most kiln
1 Trans. Am. Ceram. Soc, 18, 223 (1916).

296 JOURNAL OF THE

flues would be too short to give sufficient resistance. It was there-
fore necessary to devise a different method for providing this re-
sistance.

Apparatus.

The use of a diaphragm with an orifice, as used by the Bailey
Meter Company and others in their flow meters, used primarily
for the measurement of steam and water pressure, was thought
especially desirable. There appeared to be no reason why this
method could not be made applicable in the measurement of the
amount of gas flow in a flue system. It has been found to be
extremely accurate by the Bailey Meter Company in their meters
and preferable to the Pitot tube.

The Pitot Tube.

The Pitot tube is not suitable for measuring the flow of gases
at the low velocities usually found in flues and requires long
runs of straight flow before it will give accurate results. It is also
difficult to get an average reading with the Pitot tube, especially
when the cross section of the flue is irregular. The latter dis-
advantage would undoubtedly be met in kiln flues, for the flue
between kiln and stack is often quite short and since a straight
flue is needed on both sides of the tube, this instrument would
hardly be practical. Besides, it must be used very carefully
in order to get accurate results. Therefore, the use of a dia-
phragm with orifice in a flue, if it could be made applicable,
seemed to be the most practical.

Experimental Flue.

It was considered inadvisable to carry on the experimental
work with a flue connected to a stack because no means of con-
trolling the draft and of testing different sized flues were avail-
able. An electrically driven fan was therefore connected at its
inlet to a wooden flue (Fig. i) thus producing an induced draft.
The wooden flue was 35" X 35" X 12' in length and was made
as air-tight as possible. It had but three sides, the concrete
floor serving as a bottom. A brick flue was not considered 1
necessary — since only the drop in pressure through the orifice was |

AMERICAN CERAMIC SOCIETY.

297

neasured and all other rubbing surfaces would have but little
sffect.

A brick wall was built about nine feet from the fan inlet in the
lue. A greater distance than this was desirable but more space
was not available. The orifice in the center of the diaphragm
was made square. The brick were laid in cement mortar, and all
tracks and openings between the diaphragm and the flue were
nade air-tight.

:

S ""

%

V

i '

" f I II [j

i

1

■1

Fig. 1.

Orifices of two sizes were tried. The area of the first was equal
to about 25 per cent, of the area of the vertical section of the
flue. With a velocity of ten feet per second through the flue,
the draft gage registered a drop of 0.42 inch W. G. through this
orifice. This was considered excessive since the draft would
undoubtedly be seriously affected by much added resistance.
An orifice having an area of about 50 per cent, of the flue was then
constructed. With this orifice, the drop under the same velocity
was but o. 13 inch W. G. and at the lowest velocities was about 0.015
inch W. G. An orifice of this size was therefore adopted. The area
of the orifice (26" X 26") was a little larger than one-half (55 per

20<S

JOURNAL OF THE

cent.) the area of the flue. A sketch of the part of the flue con-
taining the diaphragm is shown in Fig. 2.

In order to make the conditions more comparable to those
actually found in flues leading from kiln to stack, the end of the
flue was partly closed. The obstruction served as the resistance
offered to the flow of the gases by the furnaces and kiln, and was
made equal to about 0.2 inch W. G. under maximum velocity
conditions. It has been found that, for ordinary round kilns,
the draft gage reaches a maximum of about 0.2 inch W. G. The

rr ^-

A - Draff Gaqe Tubes
Pifot Tube

Fig. 2. — Portion of flue showing diaphragm and positions of tubes.

.

draft gage tube "A" (Fig. 2) between the diaphragm and the er
of the flue would therefore register a pressure of 0.2 inch plus the
resistance from the end of the flue to the tube, while the draft
gage tube "A"' beyond the diaphragm would read this resistance
plus the resistance offered by the diaphragm. These gage tubes
were placed about three inches from the diaphragm on each side 1
and entered the flue through holes in the center of one side. They
projected about one inch into the flue but this distance is not
important since the gage reading was not affected by any change.

AMERICAN CERAMIC SOCIETY. 299

Measurements.
A Pitot tube was used to measure the velocity of the gas through
ie orifice for every reading taken on the draft gage. The Pitot
ibe was practical in this case for we had a straight flow of gas
id the velocity through the orifice — 400 to 1000 feet per minute
-was high enough to assure accuracy. The Pitot tube was in-
rted just behind the diaphragm (Fig. 2) and through the same
>ening used for the draft gage tube — the bent end of the Pitot
ibe reaching just to the center of the wall. The Pitot tube
lerefore gave the pressures existing in the orifice. Three read-
gs were taken across the orifice, as marked in Fig. 2 — the average
I these being taken as the average pressure in the orifice.
An inclined manometer, which could be so adjusted that from
vo to forty inches of slant along the manometer was equivalent
» one inch of vertical reading, was used. The forty-inch slope
: the manometer assured an accurate reading with low velocities.
1 converting the manometer readings to inches of water gage, they
ere multiplied by the specific gravity of the oil used (0.837).
Since one part of the Pitot tube gives total pressure and the
ther static pressure, the difference being the velocity pressure, if
ich tube is connected to the ends of the manometer tube, the
ifferential reading will be the velocity pressure. To illustrate,
ippose a reading is 3.82 and the manometer is set at a slope of
o (that is, 20 inches on the manometer being equivalent to one-
lch vertical reading). Then 3.82 -f- 20 = 0.191" vertical
fading. Multiplying by 0.837, the specific gravity of the oil,
.191 X 0.837 = 0.16" W. G. = the velocity pressure. The
elocity of the gas is then determined by means of the formula —

v - kV^,

rhere V = velocity in feet per minute,

p — velocity pressure in inches W. G.,
nd K = velocity constant.

The velocity constant K changes with the temperature. At
o° F. it is 4006. A table giving these constants is included
iter. The photograph (Fig. 1) shows the arrangement of the
ntire apparatus. The inclined manometer is shown directly
bove the word "Pitot."

300

JOURNAL OF THE

Before making any tests, readings of the Pitol tube were
checked against the tubes used with the draft gage f°r static
pressure. The manometer was also checked against the draft
gage. Each tube was inserted at the same time, the fan operated,
and the following results were obtained:

Draft gage tube

Pitot tube (static pressure)....

j On draft gage... . 0.587 inch W. G.
I On manometer... 0.591 inch W. G.

On draft gage. .. .
On manometer . . .

0.579 inch \V G.
0.582 inch W. G.

The readings of the Pitot tube and draft gage tubes varied less
than 0.01, differing but 0.008, and the manometer and draft gage
differed but 0.004. This was quite satisfactory and gave us more
confidence in the readings of the Pitot tube.

The Tests.

In making the tests, the velocity through the flue was varied —
200 feet per minute being the minimum. By partly closing the
outlet of the fan, any desired velocity could be obtained.

The results secured with the 35" by 35" flue and 26" by 26"
orifice are tabulated as follows:

Table I.

Static

pressure

before

diaphragm

(Tube A'

in Kig. 2).

Differential
reading
(draft
gage).

Manom-
eter.

Slope.

Velocity

through

orifice

in feet

per minute

0.02

O.OI7

O.4

40

370

O.048

O.O4

I .00

40

587

O.065

O.052

I.30

40

668

O.085

0.07

1 93

40

815

O. 12

O. IO

2-45

40

917

O. 13

O. 107

2-57

40

938

O.I55

O. I 16

2.78

40

977

O.I7

O. 12

2 .90

40

995

O. 19

O.I25

3 07

40

1030

Remarks.
Temperature 80 ° F.
Specific gravity of oil 0.837
Formula to figure velocity —

V = kV/>-

K at 80 ° F. = 4044
Unrestricted outlet.

The size of the flue was next reduced to 23V2" by 22 1/2" (3-67
square feet). The area of the orifice was made 55 per cent, of
that of the flue or 2.02 sq. ft. (17" X 17").

AMERICAN CERAMIC SOCIETY. 301

The results for a flue of this size are tabulated as follows:

Table

2.

Static

pressure

before

diaphragm.

Differential
reading.

Manometer.

Slope.

Velocity

through

orifice

F. P. M.

Remarks.

0.02

OOI5

O.18

20

351

Temperature 82 ° F

O.O27

0.022

O.29

20

446

K at 82 ° = 4050

O.O34

O.O28

0.33

20

476

O.O45

O.O35

O.4O

20

523

O.057

O.048

0.55

20

614

O.06

OO5

O.58

20

630

O.O9

O.062

O.88

20

776

O.I3

O.08

I .23

20

918

O.I5

O. II

I .30

20

945

O. 19

O. 123

i-55

20

1030

If the results secured by the use of the two orifices are com-
pared, it will be found that they check very closely. For in-
stance, for 370 feet per minute in the larger flue, a drop of 0.017
inch through the orifice was noted, while for 351 feet per minute
in the smaller flue, we noted a drop of 0.015 inch. Again, with
1030 feet per minute in both flues, we noted 0.123 inch drop in
one case and 0.125 inch in the other. This appears to indicate
that if the ratio of the areas of orifices and flues are kept constant,
the same drop in pressure through the orifice occurs with the same
velocity — the size of the flue being unimportant.

The accuracy of the readings was tested by the use of the
relation, "pressure or resistance varies as the square of the
velocity." The average of the two tests was taken from the
following condensed results (Table 3) giving only the drop in
pressure through the orifice for the corresponding velocity. The
higher velocities, 1100 to 1600 feet per minute, which the fan
could not produce, were calculated from the pressure-velocity
relation (Table 4). Under "ratio" is given the ratio of the first
reading of pressure to every other pressure. The second ratio
column gives the same for the velocities. Any pressure ratio
should, therefore, be theoretically equal to the square of the
corresponding velocity ratio.

3°2

JOURNAL OF THE

T..HLE 3.

Large flue.

Small fit

e.

Pressure drop.

Velocity

Pressure drop

Velocity.

0.017

370

OO15

35 1

0.04

5«7

0.022

446

O.052

668

0.028

476

O.07

815

O.O35

523

O. IO

917

O.O48

614

O. 107

938

0.05

630

O. 116

977

O.062

776

O. 12

995

O.08

918

O.125

1030

Oil

945

O.150

1 100

0.123

1030

O.178

1200

0.147

1 100

0.2IO

1300

0. 176

1200

O.243

1400

0.206

1300

O.279

1500

0.238

1400

O.318

1600

0.273
0.312

1500
1600

Table 4.

— Average

for Temperature

OF

80 ° F.

Pressure
drop.

Ratio.

Velocity through
orifice.

Ratio.

Velocity through
flue.

0.016

360

198

O.028

I

75

482

I

34

266

O.044

2

75

600

I

66

330

O.051

3

19

650

I

81

357

O.066

4

12

796

2

21

407

O.09

5

62

918

2

55

473

O. 1 107

6

92

952

2

64

523

0.115

7

19

970

2

69

533

O. 124

7

75

IO30

2

86

555

0.1485

9

29

I IOO

3

06

605

O.177

11

07

I200

3

34

660

O.208

13

0

1300

3

62

715

O . 2405

15

03

14OO

3

89

770

O.276

17

25

1500

4

17

825

0.3I5

19

68

1600

4

45

880

These ratios were plotted against each other and checked by
drawing the theoretical pressure-velocity curves on the same
chart (Fig. 3). With but two exceptions (the pressure drops of j
0.066 inch and 0.09 inch) all points fall upon the curve. This
appears to indicate that the drop in pressure through an orifice

AMERICAN CERAMIC SOCIETY.

303

is an accurate indication of the velocity of the air passing through
a. flue. The results were put in more convenient form by means
af curves so that the velocities could be determined directly from
the differential reading, providing the temperature is known

Cu

rye 5

heet

*/

f

/t

1

/

/

/

2 3 4-

Ve/ocify Pat/o
Fig. 3.

3<>4 JOURNAL OF Till-

and the same ratio of orifice and flue, as used in this Investigation,
is employed.

The temperature of the gas is an important factor in the pres-
sure developed at a certain velocity. It would appear that, as the
temperature increases and the sp. gr. of the gas decreases, the
pressure developed at a constant velocity becomes lower. vSince
the results (Table 4) were taken at a temperature of about 8o° F.,
they cannot be applied to conditions in a flue, the temperature
of which may be 15000 F. Results for the different tempera-
tures must therefore be obtained.

A very simple method of converting the pressures from one
temperature to that at another temperature was obtained by the
use of the following "Table of Constants" (Table 5)1 giving the
velocity constants already mentioned for any temperature.

In Table 5, K represents the velocities of dry air, in
feet per minute, required to sustain a column of water one inch in
height. This was obtained by using the formula

, _ -\\2 g X weight of 1 cu. ft. of water at 62 ° F.

12 X weight of 1 cu. ft. of air at temp, stated.

Small k in the fomula is the velocity in feet per second and is
multiplied by 60 to secure the values for K given in the table.
g = acceleration due to gravity = 32.16.

The ratios given in the table are those between the fan
speeds necessary for the various temperatures listed and 70 ° F.
to produce the same water gage indication.

The table is used as follows: To change any pressure, say at
80 ° F. to that at 200 ° F. under the same velocity, divide the
pressure at 80 ° F. by the square of the ratio of the velocity con-
stant at 200 ° F. to the velocity constant at 80 ° F. For example,
suppose we take the differential reading obtained for a velocity
of 1600 feet per minute through the orifice or 880 feet per minute
through the flue, which is 0.315 inch W. G. at 8o° F., the tem-
perature of the experiment. The ratio of the velocity constants

at 2000 F. and 8o° F. would be 447p_ /447p\2 = ^ = the fafr

4044 \4044/
1 This table was obtained from Mr. L. E. Eisensmith of the American
Blower Co.

AMERICAN CERAMIC SOCIETY.

305

:or by which all pressures at 80 ° F. must be divided in order
:o obtain the pressure at 2000 F. 0.315 -f- 1.22 = 0.258. The
Dressure at 880 feet per minute at a temperature of 200 ° F. would
;herefore be 0.258 inch W. G.

Table 5. — Constants.

Figured for dry air at sea level.
Barometer 29.92 inches mercury.

Temperature.

K.

Ratio.

Temperature.

K. Ratio.

4o°F.

3567

O.890

480 ° F.

5332 I

33i

30

3609

O

901

500

5389 I

345

20

3651

O

911

520

5445 1 •

359

10

3692

O

92 1

540

5500 1

373

0

3733

O

932

560

5555 1.

387

10

3773

O

942

580

5609 1 .

400

20

3813

O

952

600

5663 1 .

413

30

3852

O

961

620

57i6 1.

426

40

3891

O

971

640

5769 1 •

439

50

3930

O

981

660

5821 1.

452

60

3968

O

990

680

5873 1 •

465

70

4006

OOO

700

5925 1.

478

80

4044

009

720

5976 1 .

491

90

4081

Ol8

740

6027 1.

504

100

4118

028

760

6077 1 .

517

no

4155

037

780

6127 1.

530

120

4191

046

800

6177 1

542

130

4227

055

820

6226 1.

554

140

4263

064

840

6275 1

566

150

4298

073

860

6323 1.

578

160

4333

082

880

6371 1.

590

170

4368

090

900

6418 1

602

180

4402

098

920

6465 1

614

190

4436

106

940

6512 1

626

200

4470

114

960

6558 1

638

220

4537

132

980

6604 1

649

240

4603

149

IOOO

6650 1

660

260

4668

165

I020

6695 1

671

280

4732

l8l

1040

6740 1

682

300

4796

197

1060

6785 1

693

320

4851

213

1080

6829 1

704

340

4921

228

I IOO

6873 1

715

360

4982

243

II20

6918 1

726

380

5042

258

1 140

6962 1

737

400

5101

273

I 1 60

7005 1

748

420

5159

288

1 180

7048 1

759

440

5217

303.

I200

7090 1

770

460

5275

317

306

JOURNAL OF THE

In this way the pressures at So0 F. were converted into those at
temperatures up to 15000 F. The results lor each temperature
were plotted as shown in Fig. 4. In plotting, the two velocities
corresponding to the two points which were oil the pressure-
velocity curve were corrected in order to make the final curves

eoo 700 goo Bog

Velocity Through Flue in ft per Minute
Fig. 4.

AMERICAN CERAMIC SOCIETY. 307

►re accurate. From these curves, the velocity through the
e at any temperature can be secured from the differential draft
je reading. However, even the velocities may become mis-
ding, since the same velocity at different temperatures does not
an the same weight of gases, and the weight of the gases after
is what we must have in order to determine the rate of com-
stion we are getting in the fire boxes. A set of velocity weight
-ves for the different temperatures was therefore drawn directly
;r the differential draft gage reading curves, and from this
uble set of curves the weight of the gas may be determined
ectly from the differential reading on the draft gage. This
ight is expressed in pounds per minute per square toot of flue
:a — since this forms a basis from which the actual weight of
5 can be figured for any flue. The weights are based on dry
— since the data in the "Table of Constants" was figured on the
ne basis. A source of error is introduced here since the weight
the air varies as the relative changes in humidity. However,
\ most useful part of these curves is from about 300 ° F. and
fher. The relative humidity must become very low as the
nperature increases beyond this point and the error should not
great enough to make the results unreliable.
An illustration of the use of the curves is indicated in Fig. 4.
sume a reading of 0.132 inch W. G. at a temperature of 6oo° F.
the flue. Approximating 600 ° F. on the solid line curves, we
me down until we strike 0.132 on the differential reading scale,
we wish to determine the velocity, we drop to the abscissa and
d it to be 800 feet per minute. To get the weight we proceed
rtically in the direction of the arrow until we strike the ap-
aximate 600 ° F. curve on the broken line set of weight curves.
e then go to the left to the weight scale along the ordinate and
d it to be 30.

Conclusions.
Lack of time prevented any additional work. Since the
ferential readings at the higher temperatures are not very
ge, it would be interesting to run similar tests with a somewhat
taller orifice in order to determine the relation between the size
the orifice and the differential readings. It is very doubtful
lether the orifice should be made smaller since the size used

308 JOURNAL OF THE

produced a static- pressure drop of somewhat over <>. i inch W. (

The addition of o.i inch W. G. of static pressure to a stack woul
very likely push it to its limit in order to maintain sufficiej
velocity pressure to produce the desired draft movement.

It is interesting to note that, with induced draft, the stati
pressure gives a higher reading than the total pressure. Th
relation — "total pressure minus static pressure equals velocit
pressure" — still holds, however, since in induced draft they ai
both negative and by subtracting the static from the total pre;
sure we get a positive velocity pressure which is as it should b<
For example, it we have a total pressure of 0.4 inch W. G. and
static pressure of 0.6 inch W. G.,

then T. P. — S. P. = V. P.

— 0.4 — ( — 0.6) = +0.2.

With this, a possible method of using the draft gage to measur
velocities and weights in a flue has been found. However, th
work is merely experimental and should be tried out under actus
conditions in order to determine its practicability. It show
possibilities and might, with some alterations, be of value in th
firing of kilns. Conditions vary at every plant, but there seem
to be no reason why any one plant could not prepare a set <
readings by this method for use in noting whether or not the bur
is progressing satisfactorily.

DISCUSSION.

Mr. Davenport: I would like to ask Mr. Harrop whether
minimum static pressure observed in these tests was comparab
with the minimum observed in ordinary kiln firing. Although tl
data presented is interesting from the theoretical standpoin
it has been my experience that the very low draft pressures i
volved in periodic kiln firing render it exceedingly difficult
secure reliable readings. Mechanical engineers have about givi
up the problem of trying to read pressure differences amountii,
to only 0.04" to 0.05" of water.

Mr. Harrop: Our method of measuring the pressure involv
the use of a manometer similar to that used by the Americ

AMERICAN CERAMIC SOCIETY. 309

)wer Company. This manometer may be set at a very low
jle and is very sensitive. I realize that it is very difficult to
ure an accurate reading in this way, but to my knowledge this
nometer is the most delicate instrument available for use in
asuring the velocity pressure by the Pitot tube method. In
ard to the pressures employed, the static pressure resulting
m the resistance through a kiln at the beginning of a burn was
:en as 0.02" and was determined by an actual kiln reading.
is static pressure of 0.02" resulted from the resistance of the
ti to the flow of gases. In other words, that static pressure
uld have shown up on the kiln side of the diaphragm. As the
rn continued the static pressure increased to about 0.21" at
! highest temperature.

Mr. Davenport : I would like to add that this problem of the
;ermination of the amount of air or gas moving through kilns
in important one, but I would also like to emphasize the fact
it it it a complex one and one which has for some time baffled
: best efforts of mechanical engineers. The Thomas Electric
ter seems to offer one practically reliable method of determining
j flow of gases at very low velocity pressures; but in my own
ierience with Pitot tubes I have found, inasmuch as the varia-
n may be as much as 50 per cent, plus or minus, that an endeavor
reduce the results to absolute values below about 0.05" is
itless. I believe, however, that if work of this kind were
dertaken by ceramic engineers and an active demand were
ated for some means of measuring very low draft pressures,
it some good would come of it; but at present the Pitot tube
d slant gage are conceded by most authorities to be exceedingly
satisfactory for the measurement of low velocities.

COMMUNICATED DISCUSSIONS.

Ellis Lovejoy : Mr. Fritz has given us a very creditable piece
work and it is to be hoped that it will lead to some improvement
the important burning operations.

[ have had no respect for a draft gage for some time and the
ijority of clay workers are also lacking in respect — if I may
ige from the number of gages I have seen entirely out of com-

jio JOURNAL OF THE

mission, or running their course of usclcssncss without any i
tention. Where records are kept, the firemen have settled dol

to Tilling in the spaces in a prefunctory manner without ev
reading the gages and the data slips are added to the dusty p
in the corner of the kiln house.

A draft gage attached to a chambered continuous kiln coi
partment will show several tenths minus pressure in the eai
stages of the compartment's operation. As the fires approa
the compartment in question, the pressure will drop back to ze
and often go beyond, indicating no draft, in fact a back press*
yet the operation of the kiln moves along in spite of the readin
of the draft gage. The behavior is easily explained and t
explanation condemns the draft gage.

A differential gage will, as Mr. Fritz has shown, give us corre
draft readings which may be reduced to velocity and to volume
gas moved. The attachment of such a gage to a kiln is a proble
in itself.

Consider the chambered continuous kiln for example. Tl
natural location for the gage would be in the long flue connectii
the kiln to the stack or fan, but in the modern kiln there are tv
distinct operations carried on by the fan draft; namely, burnii
and water-smoking, each under damper control. Draft ga
readings beyond these dampers would be of no value and the
is no room between the dampers and the kiln as a rule. It wou
be impracticable to make the connections on opposite sides of t
chamber division walls because of the wide difference in tei
perature.

Down-draft kilns may have wall stacks, outside stacks in ck
proximity to the kiln, or distant stack or fan. In the first t1
instances the differential gage could only be attached to I]
stack with such adjustments as may be necessary for the differer
in altitude. The stack, and this is also true of any flue, must
large enough to accommodate a suitable diaphragm having
opening of requisite size for the kiln operation and there will be
essential difference in temperature for which correction must
made.

Besides the difference in temperature on opposite sides of 1|
diaphragm, we must take into consideration the constantly

AMERICAN CERAMIC SOCIETY. 311

ticing temperature in the kiln flue and stack, and we must also
isider that the stack and flue temperatures, will vary widely at
r given stage in successive burns, and it is not safe to assume
it, if on the sixth day of one burn we have 0.3 inch water gage
ift, this will be the proper draft for the sixth day of any other
rn.

rhe draft gage to be useful to the clay workers is not to tell
:m what has been done but to direct the work in the future,
real purpose is to enable the manager to put in the kiln house
liagram showing the burner just what the draft gage should
d at every stage in the process to the end of the burn,
rlow far the draft gage may become useful in kiln operations
aains to be seen, but it will, in kiln tests under competent
idling, determine proper kiln pressures for various wares in
eral types of kilns and thus give us the necessary data for
:ter kiln design and construction.

VIr. Fritz states that for ordinary round kilns the draft reaches
naximum of about 0.2 inch water gage.

[ have been seeking this data for some time and if he has
:urate records along this line I should very much like to have
:m. Down draft kilns vary widely in the construction of the
3r flue system and there is a greater variation in the kiln set-

g-

\ gas burning kiln with a simple floor system set with drain
: or other hollow ware is a very different matter from a closely
face brick kiln with a complicated floor flue system, burned
:h coal fired in grate bar furnace.

[t seems to me that 0.2 inch water gage is too low for a maximum
istance in down draft kilns, and I am inclined not to accept
without further proofs of actual tests under maximum re-
tance conditions.

2. B. Harrop: Regarding the statement in the impromptu
cussion offered by Mr. Davenport, relative to the imprac-
ability of Pitot tube readings under low pressure conditions,
ention should be called to two points : First, as the diaphragm
fice (where the Pitot tube was placed) was approximately
per cent, of the area of the flue, the velocity through the orifice

;i 2

AMKR1CAN CIvRAMIC SOCIETY.

was approximately double that through the Hue This gave a
all times sufficient velocities to be satisfactorily read with th(
Pitot tube. Second (as shown in Pig. ,^J, the relation bctweei
velocities determined by the Pitol tube and resistance or pressun
drop results in practically a perfect "R varies as V2" curve, whirl
in itself proves that the Pitot tube readings were amply accnrat<
for the work.

THE CLAYS OF FLORIDA.

By E. H. Seuuards.

Phe clays of Florida that are being utilized include those from
ich building brick and tile are made, the white-burning ball
^s or plastic kaolins, and fuller's earth. The production of
lding brick during 1916 amounted to 31,129 million, the value
which, including a small amount of tile and fire-proofing brick,
s $226,362. The production of plastic kaolin is limited at the
sent time to the output of three plants under the management
two companies. During 1916 five plants were engaged in
ling fullers' earth, the production from these five plants being
>ut 90 per cent, of the total production of this material in the
ited States.

n their geologic relation the clays of Florida all lie above the
gocene limestones since beneath these formations are Eocene
estones which, at least in the peninsular section of the State,
end uninterruptedly to a great depth.

lAhe fuller's earth clays lie within the Alum Bluff formation
ich, according to the vertebrate fauna obtained within recent
irs, is Miocene. The clays used in brick making vary in age
)bably from the Miocene to the Pleistocene or Recent. The
)graphic distribution of the clays likewise is irregular, although
that are being utilized are found in the northern part of the
ite — within 200 miles or less of the north line.

The Florida Common Clays.

^o clays suitable for making vitrified brick have been located
;hinsthe State. In 191 5 a series of tests of 25 samples of Florida
ys was made by the Bureau of Standards laboratory at Pitts-
rgh. The clay samples for these tests were collected by the
)rida Geological Survey. The 25 samples tested represented
ys from 21 counties in Florida. Each sample approximated
) pounds in weight and was representative, as nearly as could

|i i J( HJRNAL OF THE

be judged, of the clay <>f the locality from which it was taken.
These tests were not successful in locating any clays that would
serve in the manufacture of vitrified brick. Although previously
published in the reports of the Florida Geological Survey, the
results from a few of these 25 tests on the common clays of
Florida may be here included for convenience of reference.

Sample No. 1, Jackson County. The clay works with some
difficulty in the stiff-mud condition; water of plasticity in per cent,
of dry weight, 40.80 per cent.; warped and cracked during the
drying treatment; linear drying shrinkage, in terms of wet length,
9.95 per cent.; linear burning shrinkage in terms of dry length,
at 8500 C, 0.77 per cent.; at 1010 ° C, 4. 59 percent.; at ii3o°C,
6.60 per cent.; at 12500 C, 7.55 per cent.; color after burning,
light red at lower temperatures, changing to dark red at higher;
per cent, porosity, at 85o°C, 36.75 per cent.; at 9500 C, 40.55
per cent.; at 9800 C, 34. 30 per cent.; at 10100 C, 30.30 per cent.;
at 10400 C, 24.20 per cent.; at 10700 C, 24.00 per cent.; at 11000
C, 26. 10 per cent.; at 11300 C, 25. 15 percent.; at 11600 C, 23.65
per cent.; at 11900 C, 23.60 per cent.; at 12200 C, 23.85 per
cent.; at 12500 C, 20.70 per cent. A somewhat plastic and
sticky clay of high drying and burning shrinkage. The clay re-
tains a porous structure at 12500 C, and cannot be used
in the manufacture of vitrified ware burned in commercial
kilns. The clay may be used in the manufacture of common
and building brick.

Sample No. 2, Washington County. — The clay possesses good
working plasticity and molding behavior ; water of plasticity, 36.0
per cent.; a few cracks developed by drying; linear drying shrink
age, 9.42 per cent.; linear burning shrinkage, at 8500 C, 0.55
per cent. ; at 10100 C, 1.57 per cent. ; at 1 1300 C, 7 . 75 per cent,
at 12500 C, 8. 10 per cent. A good light buff color is developed
by burning; per cent, porosity, at 8500 C, 36.80 per cent.; al
9500 C, 35 .30 per cent.; at 9800 C, 34.75 per cent.; at 10100 C.
34.50 per cent.; at 10400 C, 30.75 per cent.; at 10700 C, 25.7c
per cent. ; at 1 1000 C, 22 . 45 per cent. ; at 1 1300 C, 19 . 70 per cent,
at ii6o°C, 17.95 per cent.; at 11900 C, 15.70 per cent.; at 1220'
C, 14.60 per cent.; at 12500 C, 12.55 per cent. A buff burning
clay of good plasticity and a relatively high drying shrinkage

AMERICAN CERAMIC SOCIETY. 315

y be used in the manufacture of buff colored face brick, al-
iigh care must be exercised in drying. The clay must be
ned above 12500 C. in order to attain low porosity.
.ample No. 3, Santa Rosa County. — The clay has good working
sticity and molding properties; water of plasticity, 28.90 per
t. ; no drying difficulties; linear drying shrinkage, 6 per cent.;
ar burning shrinkage, at 8500 C, 0.64 per cent.; at 10100
0.21 per cent.; at 1 1300 C, 1 .44 per cent.; at 12500 C, 1 . 17
cent.; burns to salmon color, changing to buff at higher tem-
ature; per cent, porosity, at 8500 C, 35. 30 per cent.; at 9500
35.80 per cent.; at 9800 C, 36.20 per cent.; at 10100 C,
86 per cent.; at 10400 C, 33. 60 per cent.; at 10700 C, 32.15
cent.; at 11000 C, 31.10 per cent.; at 11300 C, 29.55 Per
t. ; at 11600 C, 29.05 percent.; at 11900 C, 28.80 per cent.;
12200 C, 28.05 per cent.; at 12500 C, 27.30 per cent. A
t possessing good working and drying qualities but which can-
be vitrified at the burning temperatures of commercial kilns.
;t pieces burned to 12500 C. are easily cut by a knife. This
r is of value only in the manufacture of porous common build-
brick, etc.

•ample No. 4, Escambia County. — Appears to have good work-
behavior and plasticity ; water of plasticity, 2 1 . 95 per cent. ;
ing behavior satisfactory; linear drying shrinkage, 5.75 per
t. ; linear burning shrinkage, at 8500 C, 1.28 per cent.; at
o° C, 0.05 per cent.; at 11300 C, o. 26 per cent.; at 12500 C,
6 per cent.; color afterburning, light to dark red; per cent,
osity, at 8500 C, 28.65 per cent.; at 9500 C, 29.30 per cent.;
)8o° C, 28.20 per cent.; at 10100 C, 27 .90 per cent.; at 10400
27 . 75 per cent.; at 10700 C, 25.85 per cent.; at 11000 C,
95 per cent.; at 11300 C, 25 .70 per cent.; at 1 1600 C, 27.95
cent.; at 11900 C, 26 . 80 per cent. ; at 12200 C, 19. 90 per cent.;
[2500 C, 20. 20 per cent. A clay possessing good working and
ing behavior, but one which retains a porous structure at tem-
atures as high as 12500 C. The clay is suitable for common
I face brick, etc., but not for paving brick or other vitrified
re.

Sample No. 5, Walton County. — Plasticity and working proper-
good; water of plasticity, 28.30 per cent.; drying behavior

3i6 JOURNAL OF THE

satisfactory; linear drying shrinkage, 6 . 75 per cent.; linear burn
in); shrinakge, at 850° C, 0.34 per cent.; at 1010 C, 0.35 pe
cent.; al [1300 C, 2.17 percent.; at 12.50° C, 2.25 per cea
color after burning, salmon to light red; percent, porositv, at X500
C, 3360 per cent.; at 950° C, 33. 20 per cent.; at 980° C, 35.41
per cent.; at 1010° C, 34.35 per cent.; at 1040° C, 32.80 pe
cent.; at 1070° C, 30.50 per cent.; at noo°C, 29.85 per cent,
at 1 130° C, 28.60 per cent.; at 1 160° C, 27.55 per cent.; at 1 190
C, 27.80 per cent.; at 1220° C, 27.10 per cent.; at 1250° C
26.00 per cent. A red burning clay having good working an<
drying behavior, but retaining a porous structure. Test piece
burned to 1250° C. are easily scratched by a knife. This clay i
suitable for the manufacture of porous common and buildin]
brick. Not practical to vitrify in commercial kilns.

Sample No. 11, Duval County. — Fairly plastic, fairly goo
working qualities; water of plasticity, 27.4 per cent.; dries satis
factorily; linear drying shrinkage, 9.6 per cent.; linear burnin;
shrinkage, at 990° C, 0.05 per cent.; at iiio° C, 0.82 per cent.
at 1 230° C, 2.34 per cent.; at 1320° C, 3 .97 per cent.; red burn
ing; per cent, porosity, at 850° C, 28.1 per cent.; at 9500 C
26.8 per cent. ;at98o°C, 25.7 per cent.; at 1010° C, 25.8 percent,
at 1040° C, 24.8 percent.; at 1070 ° C, 24. 6 per cent.; at 1100
C, 22 .5 percent.; at ii30°C, 22.5 percent. ; at 1160° C, 20.4 pt
cent.; at 1190° C, 16.6 percent.; at i22o°C.,n.5 percent. ; at 125c
C, 7.5 per cent. A sandy surface clay of fair working and dr\
ing behavior. A decrease in porosity is noted with increase i
temperature, although it is doubtful whether a vitrified produ<
may be manufactured from this material, owing to the relative!
high temperatures necessary.

Sample No. 12, Clay County. — Very plastic, and possesses fa
working qualities; water of plasticity, 34 . 4 per cent. ; excessive dr
ing shrinkage ; linear drying shrinkage, 12.22 per cent. ; linear bur
ing shrinkage, at 990° C, 1.45 per cent.; at ino° C, 3.48 p
cent.; at i23o°C, 4.87 percent. ; red burning; per cent, porosit
at 850° C, 24.4 per cent.; at 950° C, 22.7 per cent.; at 980° (
19.6 percent.; at 1010° C, 18.7 per cent.; at 1040° C, 18.7 p
cent.; at 1070° C, 17.4 per cent.; at 1100° C, 17.3 per cen
at 1 130° C, 16.8 percent.; at 1160° C, 16.9 percent., at ii9o0(

AMERICAN CERAMIC SOCIETY. 317

per cent.; at 12200 C, 14.4 per cent.; at 12500 C, 13.8 per
A plastic red burning clay having a high drying shrinkage.

must be exercised in drying heavy pieces. This clay has a
ively low porosity at commercial kiln temperatures and at-
; a fairly dense structure.

mple No. 21, Jefferson County. — Medium plastic with fair
:ing properties; water of plasticity, 32.6 per cent.; no drying
:ulties; linear drying shrinkage, 9.77 per cent.; linear burn-
ihrinkage, at 9900 C, 0.22 per cent.; at 11100 C, 1.09 per
,; at i23o°C, 0.55 per cent.; at 13200 C, 0.49 per cent.;

burning; per cent, porosity, at 9900 C, 35.6 per cent.; at
>° C, 340 per cent.; at 10500 C, t,t, . 2 per cent.; at 10800 C,

per cent.; at 1 1 io° C, t,^ .8 per cent.; at 11400 C, 33 .6 per
. ; at 11700 C, 33.6 per cent.; at 12000 C, ^t, .7 per cent.; at
i° C, 32 .8 per cent.; at 12600 C, 34.4 per cent.; at 12900 C,

per cent.; at 13200 C, 33.5 per cent. A sandy buff burning

which retains an open porous structure at temperatures up
520 ° C. May have some use in the manufacture of soft
us common building brick.

imple No. 22, Polk County. — Medium plastic with fair work-
properties; water of plasticity, 24.9 per cent.; no drying
:ulties; linear drying shrinkage, 6.28 per cent.; linear burn-
shrinkage, at 9500 C, 0.37 per cent.; at 11000 C, 0.47 per
. ; at 12200 C, 0.08 per cent.; at 13100 C, 0.24 per cent.;
is red; per cent, porosity, at 9500 C, 35.6 per cent.; at 10100
35 .8 per cent.; at 10400 C, 35 .6 per cent.; at 10700 C, 36.0
cent.; at uoo° C, 34. 8 per cent.; at 11300 C, 33.8 per cent.;
1600 C, ^t, .8 per cent.; at 11900 C, 33.7 per cent.; at 12200
33.6 per cent.; at 12500 C, 33.7 per cent.; at 12800 C, 33.4
cent.; at 13100 C, 33.6 per cent. A sandy red burning ma-
il which retains an open porous structure at temperatures up
1320° C. May have some use in the manufacture of soft
ms common building brick.

The Florida Plastic Kaolins.

ccurrence. — The plastic kaolins of Florida present problems
xceptional interest. The formation which holds the clays is

318 JOURNAL OF THE

probably co extensive or nearly so with the physiographic typ

known as the Lake- Region of Florida. This belt of country ej
tends in the peninsula of Florida from Clay County on the nort
to near the middle of DeSoto County on the south, a distance c
about [50 miles. In width the belt varies from 10 to 30 or 4
miles. Small lakes are numerous. Their basins, as a rule-, ai
circular in outline and relatively deep with steep sides. Th
uplands are sandy and well drained. A similar type of topoj
raphy, underlaid probably by the same formation, is found i
several counties in west Florida between the Suwanee and Choctav
hatchee rivers.

The clay in this formation is intimately associated with coars
sand from which it is removed by washing. Mica is also preset
and is removed by screening. The sand in this formation
usually coarse and in places affords the sharpest and best buil<
ing sand found in Florida.

The place of the clay-bearing formation in the geologic tin
scale is difficult to determine owing to the complete absence <
fossils. It overlies the Oligocene limestones. There is also son
reason for believing that it lies at a stratigraphic level high<
than the fuller's earth beds and hence is not older than the Mi<
cene. However, inasmuch as no one of the later fossiliferoi
formations is found overlying this formation, it has not bee
possible to fix its age more definitely.

The two localities at which this clay is being worked are Edg;
in Putnam County and Okahumpka in Lake County. At Edga
4 to 10 feet of loose sand lies above the kaolin-bearing sand. Tt
top sand is coarse — containing silicious pebbles up to one-thii
of an inch across. The large pebbles are flattened and all a
rounded. The kaolin-bearing sands beneath are gray in cole
although the weathered surface is sometimes slightly iron-staine
They are said to have a total thickness of 30 feet or more and a
underlaid by a sticky, blue clay- It is reported that beneath I
blue clay a fuller's earth occurs, and that this in turn passes
a depth of about 70 feet into a scarcely indurated shell stratu:
A well put down by the Edgar Plastic Kaolin Company is 1
ported to have passed through coarse superficial sand, 10 fe
kaolin-bearing sands, 30 or more feet; sticky, blue clay with fulle

AMERICAN CERAMIC SOCIETY. 319

h beneath, about 40 feet; scarcely indurated shell stratum,
eet. The well terminated on a hard limestone at the depth of
eet, probably the Chattahoochee limestone, which is Upper
ocene in age, or possibly the Eocene limestones wh'.ch lie
beneath — the Chattahoochee formation being usually want-
along the eastern slope of the peninsula.

he kaolin in Lake County occurs under conditions similar to
e found in Putnam County. The superficial sands here, as
he Edgar mines, are coarse and contain white, silicious peb-
The kaolin-bearing sands are gray in color except where
led red with iron. A small amount of mica, found in the
in sands, is screened out in the process of washing,
he sand-clay mixture of this formation is often well adapted
oad construction, and is frequently so used, especially when
ewhat colored by iron-staining as it often is near the surface.
be Florida kaolin-bearing formation is plainly sedimentary
rigin and represents, as indicated by the rather coarse sand,
latively near-shore accumulation of material. The associa-
of the finely divided clay with the coarse quartz sand and the
1 is one of the problems of this formation.

ining. — In mining the Florida kaolin the overburden, which
ists of a few feet of sand or iron-stained sand and clay, is
oved — usually by the hydraulic process. The clay itself is
i chiefly by suction pumps which are carried on a floating
ge. The pits are first of all opened to the water table level,
•h lies at a moderate although varying depth, depending upon
Lopography. The dredges are then floated on the water that
mulates in the pit. The dredge itself carries a relay station,
some of the dredges this relay pumping station, eorrespond-
:o the sump-hole of ordinary hydraulicing, is itself submerged,
pump removing the clay from the sump-hole has a capacity
tly greater than the pump which brings the clay to the sump-
Thus, notwithstanding the fact that the process goes on
below the surface of the water, no part of the clay is lost at
relay station because all that goes into the bin is removed
nee — the suction of the pump preventing the escape of the
ing particles of clay.

320

jorkXAI, OK THIv

Fig.

AMERICAN CERAMIC SOCIETY. 321

rom the relay pumping station on the dredge the material
ses through the washing station where the coarse sand and
r balls are removed to settling tanks. From the settling
ks the clay is again pumped into compressors. After the
sss of moisture is removed by compression, the clay is further
ficially dried for shipment,
ill of the plastic kaolin produced in Florida is shipped out of

State and is used chiefly or entirely in mixing with other
fs where it is of value because of being plastic as well as white
ning and refractory. Chemical analyses of this clay have been
viously published and are accessible. On the map, which ac-
lpanies this paper, is indicated the location of the plants in
rida which are producing this clay, and also the location of

brick clay and fuller's earth plants of the state. The bibliog-
hy which follows includes references to some of the relatively

papers that have been published on the Florida clays.

Ries, Heinrich, "Clays of the United States East of the Mississippi River,"
S1. Geol. Survey, Prof. Paper No. n, pp. 83-85 (1903).

Matson, George C, "Notes on the Clays of Florida," U. S. Geol. Survey,
7. 380, 346-357 (1909).

Sellards, E. H., "Mineral Industries and Resources of Florida," Fla.
:e Geol. Survey, Sixth Annual RepL, 1914, pp. 23-35.

Memminger, C. G., "Florida Kaolin Deposits," Eng. Mining J., 57,

(1894).

Florida State Geological Survey.

AN ATTEMPTED HEAT BALANCE OF A CONTINUOH
KILN OF THE CHAMBER TYPE.

By C. TrBISCHBL and II. S. ROBERTSON, Schenectady, N. Y.

The available ceramic literature at the present time shows
great scarcity of heat-balance data. The relatively long te
periods necessary in securing authentic results has tended towai
discouraging the collection of such data, the value of which hi
long been recognized. It is with the idea of "helping along tl
cause" which has been sustained by Bleininger,1 Gelstharp2 ax
Harrop,3 that the authors have decided to present this data
the Society. The tests were conducted on one of the kilns at tl
Alton plant of the Alton Brick Company, Alton, 111.

The Kiln. — The kiln was of the chamber, moving fire typ
having fourteen chambers arranged in one continuous row. Ea<
chamber received its fire from one side only, working on the dowi
draft principle, the flames coming up over a bag-wall, spreadit
over the chamber and passing out through the floor. That ha
of the floor nearest the bag-wall was closed, the other half w;
open, thus providing for a good distribution of heat over the e:
tire chamber. A connecting flue for carrying the hot combu
tion gases from chamber to chamber extended the full leng
of the kiln beneath the floor level. The waste combustion gas
and waste water-smoking gases were led from each chamber I
auxiliary flues which connected with the main waste-gas and ma
water-smoking flues through bell dampers, all being undergrour
These main flues were parallel with the kiln and each was pi
vided with a fan for creating draft.

Water-smoking was accomplished by drawing the gases from t
cooling chambers behind the fires through a sheet-steel flue si
ported about ten feet above the top of the kiln. (This is kno1

1 Trans. Am. Ceram. Soc, 10, 412 (1908); 11, 153 (1909).

2 Ibid., 12, 621 (1910).

3 Ibid., 19, 216 (1917); This Journal, i, 35.

AMERICAN CERAMIC SOCIETY. 323

; the short-circuiting flue.) The gases pass down through the
lambers which are water-smoking and out into the main water-
noking flue.

The fuel used, producer gas, was supplied through a large main
ie extending the full length of the kiln, below the floor level
id on the opposite side from the main water-smoking and main
aste-gas flues. Distributing flues, one for each chamber, rose
jrpendicularly to the top of the kiln, at which point each led off
; right angles into a horizontal flue which extended across the
In, parallel to the axis of the chambers. These horizontal flues
ere integral with the kiln brick work and placed between the
lambers. The gas was then delivered by small ducts to the
iot of the bag-walls.

The Gas Producers. — The gas producers were in two batteries
t four each, being of the simple or carbon-monoxide type,
hey were rectangular in plan with a single hopper at
le top for feeding the coal. Two holes were provided in the
de walls, through which a poker could be thrust for stirring
p the fuel bed. No seal was used, the coals resting on hori-
mtal grates at the bottom. The firing interval was fifteen
linutes and the grates were cleaned every eight hours.

The Ware. — The wares burned were shale paving block and hard
urned shale building brick, both maturing at about cone 4, or, to
1 exact, 11200 C. The pavers were wire-lug-cut and the build-
's were side-cut. The percentage of No. 1 ware produced was
igh, showing a good distribution of heat from top to bottom.

Testing Apparatus Used.

(a) Gas Testing Apparatus. — This consisted of a Burrell
)mplete gas analysis outfit of the portable type, provided with
ipettes for the determination of CO2, CO, 02, and unsaturated
pdroearbons by absorption and H2 and CH4 by combustion,
oth the flue-gas and producer-gas were analyzed with this outfit.

(6) Junker's Gas Calorimeter. — It was planned to use this
jparatus in obtaining the heating value of the producer-
is, but owing to adverse conditions the results could only be
sed as checks, the heating value being obtained by calculation.

.;.•! JOURNAL OF THE

(c) Thermocouples. A noble metal couple was used for high
temperature work and two base metal couples for obtaining 1 lit
temperatures in the producer j;as and waste COjnbustion-gas flues.

(d) Gas Collecting Apparatus. — In order to obtain true gas
samples from the various flues and the interior of the burning
chambers, pieces of 3/4" iron pipe were used. These varied in
length from 6 to 9 feet, were closed at one end and bushed down
at the other to take a 12-in. length of '// pipe. Holes were drilled
at random along the length of the 3/4" pipe, those nearest the bot-
tom being larger than the ones at the top. The gas sample warn
drawn through these tubes into a large pressure tank. This
pressure tank was designed similar to the large tanks used by gas
companies. Its cubic capacity was about 2T/4 cubic feet. In
operation, the lower half was filled with water which had been
saturated with the gas to be sampled. The upper half was then
placed in position and lowered until the water started to flow
from the intake pipe in its head. The tank was then connected
to the sampling tube by a long piece of rubber hose. Then by
raising the upper half of the tank a vacuum was created, the water
forming a seal, and the gas flowing into the tank. By emptying
and refilling the tank two or three times, a sufficiently pure sample
was obtained. Then, if it was desired to use the Junker calorim-
eter for obtaining the heating value of the gas, all that was neces-
sary was to connect the full tank to the calorimeter, the weight of
the upper half of the tank being sufficient to maintain the proper
pressure.

(e) Thermometers. — A centigrade thermometer was used in
obtaining the temperature of the gas in the waste water-smoking
flue and of the atmosphere. Wet- and dry-bulb thermometers
w^ere not used for obtaining the humidity of the air, as the test
was carried on during a rain-storm and heavy mist, so that the
relative humidity was 100 per cent.

The Heat Balance.

In making a heat balance on a producer-gas fired continuou
kiln the following factors must be determined :

AMERICAN CERAMIC SOCIETY. 325

A. Heat introduced as fuel.

B. Heat lost by un-burned carbon in ashes.

C. Heat lost in the producer.

D. Heat used in the burning of the ware.

E. Heat lost in the combustion-gases.

F. Heat lost in water-smoking gases.

G. Heat taken up by the kiln and lost by radiation.

In order to obtain the data for the above factors the following
>bservations, analyses, etc., were made:

Methods Used in Obtaining Data.

1. Tons of Coal Used. — The weight of coal required to charge
he hopper of the producer to each of three different depths was
ietermined and subsequent charges were noted as 1/i, V2 and
ull hoppers. In each case the depth of charge was determined
)y measurement and recorded on a firing sheet. This process
vas continued for a complete cycle of the kiln, which covered a
jeriod of four hundred and seventy-eight hours. Knowing the
veight of coal corresponding to each of the three depths, the total
imount of coal used during the period was calculated.

This method must necessarily depend upon the intelligence
)f the firemen and the interest which they take in the work.
Small errors will be of a compensating nature, so that large
irrors need only be guarded against.

2 and 3. Heating Value and Analysis of Coal. — An average
:ample of the coal was taken over a period of seventy-two hours,
iach time the fireman fired his producer he threw a small quan-
ity of coal on a pile. At the end of the period this pile was quar-
ered to a small pile and gathered in a sampling can. This sam-
>k was analyzed by the Department of Applied Chemistry of
he University of Illinois under the direction of Mr. J. M. Lind-
en.

4. Analysis of Producer Ashes. — Each time the fireman
•leaned the grates, he threw a small quantity of ashes onto a pile.
Phis pile was quartered down and the sample gathered in a sam-
)ling can. This sample was also analyzed by the Department of
Applied Chemistry, Univ. of 111.

326 JOURNAL OF THE

5 and 6. Analysis of Producer-Gas and Flue-Gas. These

analyses were made- with a Burrell apparatus.

7. Temperatures of Gases in Producer-Gas and Waste-Gas
Tunnels. A base-metal thermocouple was inserted into the
line through an opening in the top and readings taken in milli-
volts with a portable Leeds and Northrup potentiometer equipped
with compensation coils for the temperature correction of the cold
junction of the thermocouple. The thermocouples were cali-
brated and the millivolt readings were transferred to tempera-
ture readings by interpolation from the calibration curves.

8. Maximum Burning Temperature. — A platinum-platinum
rhodium thermocouple was inserted through the crown into the
chamber on high fire. Readings were taken with a potentiom-
eter.

9. Humidity of the Air. — As stated before, the relative humid-
ity was 100 per cent., due to the fact that rain and heavy mist
fell throughout the test.

10. Humidity of Waste Gases. — Due to the inaccessibility of
the top of the waste-gas stack and the large volume of sulphur-
gases coming from the stack, this determination was neglected.

11. Tons of Clay Burned. — The number of bricks or blocks
set in each chamber was obtained from records in the company
office. Ten blocks and ten bricks were weighed, to determine
the average weight of each burned block or brick. From these
weights and the number of blocks or bricks set in the kiln the
tonnage of the kiln for the entire cycle was calculated.

12. Heating Value with Junker's Calorimeter. — This method
is a direct method for obtaining the heating value of a gas. A
known volume of gas is burned and allowed to heat a known
weight of water, the temperature rise in the water being observed.
From this data the heating value is calculated. Due to the fact
that conditions were not favorable for the accurate working
of the calorimeter — there being drafts blowing over the work-
table, and a variable gas supply — these results were used only
as a check, the calculated heat value being used throughout the
work.

AMERICAN CERAMIC SOCIETY. 327

Method of Calculation,
'he method of calculation employed followed the general
hod set down by Bleininger, Gelstharp and Kratz.1 In
nfic cases the method was altered to suit the conditions of
data used.

'he items to be calculated in the determination of the factors
sssary for a complete heat balance such as this are:
l. Heat Introduced as Fuel.

1 . Number of pounds of coal fired per ton of burned clay.

2. Number of B. t. u. in coal used per ton of burned clay.
1. Heat lost in un-burned carbon in ashes.

'. Heat lost in gasification.

1. Cubic feet of gas per pound of dry coal.

2. Heating value of gas per cubic foot.

3. Sensible heat in gas per pound of dry coal.

4. Total heat in gas per pound of dry coal.
). Heat used in burning the ware.

1. Heat used in driving off hygroscopic water.

2. Heat used in dehydrating the clay.

3. Heat used in heating up clay substance.
I. Heat lost in the combustion-gases.

1. Cubic feet of flue-gas per cu. ft. of producer-gas.

2. Cubic feet of flue-gas per pound of dry coal.

3. Sensible heat in flue-gas per pound of dry coal.
\ Heat lost in water-smoking gases.

1. Cubic feet of gases per pound of dry coal.

2. Sensible heat of gases per pound of dry coal.
x. Heat absorbed by kiln and lost by radiation.

l. Heat Introduced as Fuel. — This was calculated on the
is of pounds of coal per ton of ware burned. The total weight
:oal used was divided by the total tonnage of the kiln. This
e the pounds of coal used per ton of ware. Multiplying this
dt by the heating value of the coal per pound as fired gave
heat introduced.

>. Heat Lost in Unburned Carbon in Ashes. — The percentage
mburned carbon in the ashes was obtained from the analysis

1 Garland and Kratz, "Tests of a Suction Gas- Producer," Eng. Exp. Sta.,
v of 111., Bull. 50.

328 JOURNAL OF THE

of the ashes. Since the earthy matter in the ashes correspoa

to the "ash" of the eoal, the unburned carbon in the ashes I
pound of eoal is readily obtained. The total ashes per pound
coal equals ioo per cent, divided by the percentage of ash in tl

ashes and multiplied by the percentage of ash in the coal. Thi
multiplied by the percentage of carbon in the ashes, gives the pe
centage of unburned carbon in the ashes in terms of one pom
of coal. Multiplying this result by the heating value of carb(
per pound gives the heat lost.

C. Heat Lost in Gasification. — Since the basis of the efficien<
calculations is one pound of dry coal, the volume of produce
gas per pound of dry coal must be first obtained. The weigl
of carbon combined in CO2, CO, and CH4 per cubic foot of pr
ducer-gas is calculated from the per cent, compositions. Tl
weight of carbon burned on the grate per pound of coal is tl
difference between the per cent, of carbon in the coal and the pe
cent, of unburned carbon in the ashes in terms of one pound
coal. This difference, divided by the weight of carbon per cuh
foot of gas, gives the volume of gas per pound of dry coal.

The heating value of the gas per cubic foot is calculated fro
the per cent, composition and heating value of the combustib
constituents. In obtaining the total heat of the gas per cub
foot, it is necessary to add to the heating value, the sensible he
at the temperature at which the gas enters the kiln. This is 0
tained from the per cent, composition, the specific heat of I
constituents, and the temperature. Multiplying the total he
per cubic foot by the cubic feet per pound of coal gives the he
ing value of the gas per pound of coal. This value, subtract
from the heating value of one pound of coal, gives the heat 1(
by gasification.

D. Heat Used in Burning the Ware. — In this calculation it
assumed that the clay contains 2 per cent, hygroscopic wa ,
and three per cent, of chemical water. The total heat absorb
is, therefore, the heat used in driving off and turning into ste;
the hygroscopic and chemical water plus the heat used in bu
ing the clay substance. The latter is obtained by multiply

I

AMERICAN CERAMIC SOCIETY. 329

weight of clay by the specific heat of clay and multiplying
result by the temperature rise.

i. Heat Lost in Combustion-Gases. — The volume of flue-gas
cubic foot of producer-gas is obtained by dividing the weight
carbon in the producer-gas per cubic foot by the weight of
)on in the flue-gas per cubic foot. This result, multiplied by
cubic feet of producer- gas per pound of dry coal, gives the
lme of flue-gas per pound of dry coal. From the per cent,
iposition of the flue-gas, the specific heats of the constituents
the flue-gas temperatures, the temperature of the flue-gas,
the volume of flue-gas per pound of coal, the heat lost in the
ibustion-gases is readily calculated.

. Heat Lost in Water-smoking Gases. — Due to the fact that
water-smoking gases left the flue at approximately atmo-
eric temperature, this loss was negligible and was not taken
> account.

r. Heat Absorbed by Kiln and Lost by Radiation. — As the
:ors which enter into this calculation are so variable and the
a so difficult to obtain, it is determined by the difference be-
en 100 per cent, and the summation of the calculated factors
he heat distribution. It is plainly evident that radiation will
small in this type of kiln — due to the great thickness of the
Is, the small areas subjected to high temperatures at any
time, and the high velocity with which the gases are moving
DUgh the kiln.

Data.

Coal used, as obtained from data sheets and recorded

by producer firemen '. . 524,728 lbs.

B. T. U. value of coal per pound by combustion in oxygen
bomb calorimeter and furnished by the University of
Illinois, Department of Applied Chemistry. Dry basis... 12,630

As received ... 11,175

Tonnage of burned clay in the 14 chambers 1743.0 tons

Coal used per ton of burned clay 301 .0 lbs.

Coal used per 1000 blocks 1400.0 lbs.

Coal used per 1000 builders 834.0 lbs.

Average weight of blocks 9.3 lbs.

Average weight of builders 5.7 lbs.

330 JOURNAL OF THE

Average temperature of producer-gaa Hue 7080 F.

Average temperature of waste-gas flue 406 ° F.

Average temperature of water-smoking flue- 6o° F.

Average temperature of atmospheric air 6o° F.

Maximum temperature of burning 1 1200 C.

Analysis of Coal.

Proximate. As

Ultimate. Dry. received.

C 69 . 90 Fixed carbon 41 . 30 37 . 30

H2 4-95 Volatile matter 44-48 38.58

02 10.29 Moisture 0.00 n-53

N2 0.62 Ash 9.52 8.43

VS 4-71 vSulphur 47i 4"7

Ash 9-52

Analysis of Ashes.

C = 22.24
Ash = 77.78

Average Gas Analysis (samples taken over a period of
hrs.).

Producer-gas. Flue-gas.

C02 = 4.3 C02 = 3-5

CO = 24.4 02 = 17. 1

CH4 = 1.5 N2 = 79-5

H2 = 7-4

02 = 0.6

N2 = 62 . 1

Calculations.

Calculation of Producer Efficiency. — For each cubic meter

1 2
CO2, CO and CH4 kilograms of carbon are required.

■ 22 . 4

1 2

0.043 X = 0.023 Kg. of C in CO2 per M3 of gas.

22 .4

1 2

0.244 X = o. 131 Kg. of C in CO per M3 of gas.

22 .4

12
0.015 X - — = 0.008 Kg. of C in CH4 per M3 of gas.
22.4

o. 162 Kg. of C in one M3 of gas.

o. 162 X 2 .2

35-25

= 0.0101 lb. of coal per cu. ft. of gas.

AMERICAN CERAMIC SOCIETY. 331

Carbon per lb. of coal from analysis = 0.6990 lb.

Jnburned Carbon in Ashes per Lb. of dry coal = X

77.76

>952 X 0.224 = 0.0272 lb.

_ 1U f 1 1 O.699O 0.0272

/ olume of producer gas per lb. 01 dry coal = — =

0.0101

5 cu. ft.

Heating Value of Gas.

).244 X 323.5 = 79.9 B. T. U. in CO per cu. ft. of gas.
>.074 X 326.2 = 24. 1 B. T. U. in H2 per cu. ft. of gas.
>.oi5 X 1009.2 = 15. 1 B. T. U. in CH4 per cu. ft. of gas.
Total, 119. 1 B. T. U. per cu. ft. of gas.
Sensible Heat in Gas.
>43 X 0.0257 X 708 = 0.82 B. T. U. in C02.
)6 X 0.0177 X 708 = 12.00 B. T.U.inCO,02,CH4,H2andN2.
12.82 B . T. U. per cu. ft. of gas.

Total Heat of Producer-Gas per Pound of Dry Coal.

>.82 X 66.5 = 852 B. T. U. sensible heat

).i X 66.5 = 7920 B. T. U. at atmospheric conditions

8772 B. T. U. total heating value of gas per lb.

dry coal.

772

- X 100 = 69.4 per cent. = efficiency of the producers.
S30

Per Cent. Heat Lost by Unburned Carbon in the Ashes =

3273 X 14600 X 100

— = 3 • x5 Per cent.

12630

Per Cent. Heat Lost by Gasification in the Producers =
)— (69.4 — 3.15) = 27.45 percent.

Thermal Efficiency of the Kiln.

.ximum temperature of burning = ii20°C.

erage atmospheric temperature = 200 C.

nperature rise of clay in chambers = noo° C.

at of dehydration of clay1 = 200 G. cal. per G. of water.

ent heat of hygroscopic water1 = 476 G. cals. per G. of water.

1 Clay is assumed to contain 2 per cent, hygroscopic water and 3 per cent,
mical water.

332 JOURNAL OF THE

Hygroscopic water leaves at 200 ° C.
Dehydration temperature of clay = 650 ° C.
The specific heat of clay = o . 200.

Distribution of Heat in Clay.

Hygroscopic water 0.02 X 180 X 1 = 3.6 Kg. cals.

0.02 X 478 = 9.5 Kg. cals.

Chemical water 0.03 X 0.02 X 650 = 3.9 Kg. cals.

0.03 X 200 = 6.0 Kg. cals.

Clay 0.97 X 0.200 X 1 100 = 213.4 Kg. cals.

Total heat required by 1 Kg. of clay = 236.4 Kg. cals.

236 .4 Kg. cals. = 425 B. T. U. per lb. of clay.

1 ton of clay requires 2000 X 425 = 850,000 B. T. U.

Heat introduced as coal per ton of clay = 301 X n 175 =

3,363,675 B. T. U.

Efficiency of kiln in terms of coal used at the producers =

850,000
— — — — =25.3 per cent.
3.363.675

Efficiency of Kiln in Terms of Dry Coal. — X 25.3 = 22.3.

12630

Percentage of Heat Used in Burning Ware = 25.3 per cent.

Heat Lost in Waste-Gas Flue.

o. 244 X 1 = o. 244 cu. ft. of CO2 from CO per cu. ft. of gas.
0.015 X 1 = 0.015 cu. ft. of CO2 from CH4 per cu. ft. of gas.
0.043 X 1 = 0.043 cu ^. of C02 from CO2 per cu. ft. of gas.

0.302 cu. ft. of CO2 in combustion-gas from 1 cu.

ft. of producer-gas.
o . 035 per cent, of CO2 actually in combustion-gas
from analysis.

— — =8.63 cu. ft. of combustion-gas from 1 cu. ft. of producer-
o 035

gas.
8.63 X 0.036 X 66.5 = 20.1 cu. ft. C02 per lb. of coal.

AMERICAN CERAMIC SOCIETY. 333

.63 X o . 1 7 1 X 66 . 5 = 92 . 2 cu. ft. 02 per lb. of coal.

• 63 X 0.795 X 66.5 = 456.0 cj. ft. N2 per lb. of coal.

.074 X 0.5 X 66.5 = 2 . 5 cu. ft. of water per lb. of coal from H2.

.015 X 2 X 66. 5 = 2.0 cu. ft. of water per lb. of coal from H2.

aoo X 0.03 = 60 lbs. of water per ton of clay.

5o
- = o . 20 lb. of water per lb. of coal from the clay.

31

22 4
.20 X — — X 1 .31 X 27 = 8.8 cu. ft of water per lb. of coal,
18

from the clay.

13 .3 cu. ft. of water in gas per lb. of coal.

20. 1 X 0.0248 X 406 = 201 B. T. U. in C02 per lb. of coal.
554.2 X 0.0175 X 406 = 3940 B. T. U. in 02 and N2 per lb.

of coal.

*3-3 X 0.022 X 406 = 119 B. T. U. in H20 per lb. of coal,
btal heat lost in flue-gas = 4260 B. T. U. per lb. of coal.

= 33 • 7 Per cent, heat lost in flue-gas.

12630

Heat Absorbed by Kiln and Lost by Radiation. — 100 —

5.15 + 27.45 + 22-3 +33-7) = 13 -4 per cent.

Summary.

Per cent.

Heat lost by carbon in the ashes 315

Heat lost by gasification in the producer 27 . 45

Heat used in burning the ware 22.3

Heat lost by waste-gas 33 7

Heat lost in water-smoking gases 0.0

Heat lost by radiation and conduction 134

Total, 100.00

While we are skeptical as to the exactness of the data obtained
id the reliability of some of the constants used in the calcula-
ons, we feel that a method is given which, with variations to suit
le conditions demanded, is very well suited for the obtaining
I data for a heat balance on any type of movable fire continuous
iln.

In conclusion, we wish to express our indebtedness to Mr.
. J. Hoehn for the standardization of the couples used, Mr.

334 JOURNAL OF THE

R. K. Ilursh for suggestions, and Mr. K. B. Rodgers, of the- Alton
Brick Co., for his hearty cooperation in the work.

COMMUNICATED DISCUSSIONS.

C. B. IIarrop: In reading over this paper, I have been par-
ticularly interested in the figures covering kiln efficiency and heal
used in burning the ware.

In calculating the "efficiency of the kiln in terms of coal used

at the producers," the ratio - is used giving an efficiency

3.363.675
of 25.3 per cent. The numerator is calculated on the basis of
1 .02 tons of green clay (including its hygroscopic water), instead
of 1 .052 tons of green clay (which equals 1 ton of burned clay),
while the denominator is calculated clearly on the basis of 1 ton
of dry clay. Recalculating this on the basis of 1.052 tons of
green clay, as equalling 1 ton of burned clay, gives an efficiency
of 26.1 per cent, instead of 25.3 per cent. Furthermore, this
figure should not be called the efficiency of the kiln, but rather
the efficiency of the kiln and producers, as part of the inefficiency
is due to the particular producers employed and is not a func-
tion of the kiln.

These are small matters, however, when there comes up the
difficult question of what should constitute the value of the J
numerator in the efficiency ratio. Mechanical efficiency is the I
ratio of out-put to in-put. Thermal efficiency should probably
be regarded as the ratio between the amount of heat put to use-
ful purpose (or actually doing necessary work) and the total heat
applied. The question immediately arises as to what consti-
tutes this useful purpose. If a chimney were being employed
to supply kiln draft, the stack gases should have a temperature
of from 400 ° to 600 ° F. If the stack temperature were reduced,
by carrying the products of combustion through additional
chambers, to a temperature just high enough to create the requi-
site draft, the heat then contained in the stack gases should cer-
tainly be considered as useful, for without it the kiln could nol
operate. On the other hand, any heat in excess of this would tx
wasted.

AMERICAN CERAMIC SOCIETY. 335

Even if the water-smoking gases left the kiln at 500 ° F. but
completely saturated, this heat would represent a necessary work
and should not be considered a waste or loss. On the other hand,
if the water-smoking gases were not saturated, the heat repre-
sented between their temperature and the dew point would be
a loss, while the difference between this loss and the total amount
of heat carried would be useful.

Referring again to the calculation showing that the "percentage
of heat used in burning the ware equals 25.3 per cent.," the
data taken during the test show that 301 pounds of coal were
burned at the producers for every ton of ware out-put. In other
words, considering units and thinking of only the fundamentals
of the burning operation — we set 1 . 05 tons of green ware at
atmospheric temperature; we fire 301 lbs. of coal, then we re-
move one ton of burned ware at atmospheric temperature. What
heat work has been accomplished? Merely that which is incident
to the elimination of hygroscopic water, dehydration of the clay,
and certain other pyro-chemical changes within the clay mass.
If the ware had been removed from the kiln at the maximum
temperature of 11200 C, then the author's calculation would
have been proper — but instead of being removed from the kiln
this heat was transferred by radiation, conduction, and convec-
tion (principally the latter) to the preceding chambers.

To state it differently the calculation shows that 25 .3 per cent,
of the heat in 301 lbs. of coal is used in burning one ton of the
ware and the difference (100 — 25.3), 74.7 per cent, is wasted or
lost. If this is true, and 25.3 per cent, of all the heat from the
producers finds its way into the ware at its highest temperature
in each and every chamber, then what becomes of the heat as
these chambers of ware cool? It cannot be wasted or lost — ■
because the 74 . 7 per cent, of heat from the producers covers that.
The error is apparent and very large.

Using the author's figures, the "percentage of heat used in
burning the ware" should be based on the heat used in driving
off the hygroscopic water and dehydrating the clay and should
not include any sensible heat in the ware as a result of tempera-
ture attained in the kiln.

336 JOURNAL OF THE

There is our other probable source of error in this same
connection, namely, the failure to include the heat required
to evaporate the mechanical water in the ware when set. This
usually amounts to from i to 3 per cent, or more -even after hav-
ing passed through a good drier.

C. TrEischel: I have very carefully noted Mr. Marrop's
discussion of this paper and, after re-checking the data, find
that there is an error in the kiln efficiency calculation as he has
shown.

With reference to the remainder of the discussion, however, I
am obliged to take exception. Efficiency is the "ratio between
out-put and in-put." Thermal efficiency, we agree, is the "ratio
between the heat put to necessary work and the heat in-put."
But what is the necessary work with reference to this particular
kiln? The draft is produced mechanically. Is it necessary for
a draft fan to operate on a gas mixture at 400 ° F. in order to
produce draft? If not, what is the temperature? I have seen
several kilns having mechanical draft in which the temperature
of the waste-gases was very near atmospheric temperature. Un-
fortunately, there is no data available to show whether or not
these kilns were more efficient than the one described in our paper.
However, the operators of two of these kilns have told me that
when the waste-gases leave at a higher temperature, their kilns
burn more coal. If this be true, and I have every reason to be-
lieve it is true, then, if the same facts will apply to the kiln in
question, all the heat leaving from the waste-gas flue above that
leaving at atmospheric temperature is a loss.

There is practically no loss in the water-smoking operation —
due to the fact that the gases leave at a high humidity and at
atmospheric temperature.

Now then, are we much in error when we say that this kiln is
only 25 .3 per cent, efficient? The heat lost by gasification in the
producers and in the un-burned carbon in the ashes must be called
a loss. The heat leaving the kiln with the flue-gases must be re-
garded as lost. There remains then only the loss through radia-
tion, etc., which is not unreasonable when one considers the thick-
ness of the walls, crown insulation, and all other details of con-

AMERICAN CERAMIC SOCIETY. 337

struction of this type of kiln. Disregarding small errors, the
general method of procedure seems correct.

Then, looking at the proposition from a practical standpoint,
the kiln uses 201 pounds of coal in producing one ton of ware —
25.3 per cent, of the fuel being used in the actual burning. As
this ton of ware cools, the heat is drawn off and put to some other
purpose, such as heatiug the ware in the preceding chambers —
the operation being continuous. If this were not done the 301
pounds of coal might become 500 lbs., or even 600 lbs., per ton
of ware, in which case the efficiency would be materially reduced.

NOTES ON THE LABORATORY TESTING OF SILICA

BRICK.

Hv R. J. Montgoubry an» I,. K Opmcb, Pittsburgh, Pa.

The importance of the accurate testing of silica brick is rc-
ceiving more and more attention as is indicated by the number of
articles written on the subject and the amount of work being done
at the various ceramic laboratories. This discussion is intended
to cover certain tests that are now quite commonly employed in
determining the quality of silica brick, and in order to present the
experience of the authors as an addition to the data so far pub-
lished.

The recent unusual demand for silica brick has placed upon the
market a number of new brands and has led to the investigation
of various kinds of raw materials in order to determine their
suitability for the manufacture of silica brick. In addition
to the typical ganister rock, a number of other silicious materials
such as chert, sand rock, and materials which grade from a fair
ganister to quartz sand, loosely bonded together, are being tried.
At present, there are at least two grades of silica brick on the
market. The line between these grades is not sharp and their
classification would vary with the purposes for which the brick
are used. In a classification based upon the fusion point, we
could say that Grade A would fuse at cone 32 or above and
that Grade B would fuse below cone 32.

In the testing of silica brick, it must be remembered that they
differ from fire clay brick in that they are composed almost
entirely of one substance, silica. Therefore, the properties of the
silica used govern the properties of the silica brick — this being
kept in mind in the following discussion of certain important
points in the testing of silica brick:

1. Raw Materials. — The recognized ganister rocks are known
to be capable of being manufactured into silica brick of high
quality, while the quality of brick manufactured from the other

AMERICAN CERAMIC SOCIETY. 339

kinds of silica rock is still open to question. Undoubtedly, for
many purposes, the one class of material will answer as well as the
other — if the brick are properly manufactured and the raw silica
is of uniform quality. A picked sample of sand rock or chert
may have a silica content of 96 per cent, or above, but in handling,
clay and other impurities are likely to be included and the re-
fractoriness of the brick may thus drop below cone 32. Another
point of importance to be considered in selecting the silica is the
physical strength of the brick after burning. Ganister rock crushes
down into splintery fragments and the brick has a higher crushing
strength in the cold condition than does a brick from sand rock which
crushes to rounded grains. In manufacture, the sand rock brick
are burned from 3 to 4 cones lower than are the ganister brick.
It is possible that a sand rock brick, in which the impurities
present act as a bond, although having a lower fusion point would
give a brick as strong physically as one made from crushed ganister
rock, providing it were burned to the proper cone. This, however,
would increase the physical strength at the expense of a lower
softening point.

2. Visual Inspection. — Field and factory inspection can do
much to insure uniform quality after the raw silica has been
found satisfactory. Manufacturing defects will alwavs be present.
Cracked and spongy brick are easily rejected and under-burning
may be detected by linear measurements and tapping with a
hammer. After the desired expansion in burning has been
established, careful measurement will detect any serious under-
burning. The measurement test, however, is not entirely satis-
factory and should be supplemented by laboratory tests.

3. Chemical Analysis. — The chemical analysis is of great
assistance in determining the purity of the raw material used and
should be considered in connection with an actual inspection of
the silica deposit or the raw material at the plant. For control
work the value of chemical analyses is doubtful. To secure
reliable analytical results the services of a high-grade chemist are
necessary and this makes the analytical work very expensive.
Tests other than the chemical analysis may be used as a check
on uniformity.

340

|< MkXAi. OF THE

4. Deformation Test. The ordinary deformation test,
using standard pyrometric cones for comparison, is of value.
A silica brick o! the best quality, made from pure ganister, will
not soften below cone 32 ' 2 or 33. Practically all of the brick
made from chert or sand rock have a softening point of from
cone 31 to 32. The softening temperature of a highly silicious
material is not sharply defined and the determination is difficult
owing to the high viscosity of the softened material.

5. Load Test. — The load test for silica brick, as recommended
by the American Society for Testing Materials, specifies a load of
25 lbs. per sq. inch at a temperature of 15000 C. Very few
failures have been recorded under these conditions and it is a
question whether the test is of value unless for the testing of new
brands of brick. When failure occurs, it is due to shearing or a
rupture of the bond and not to a softening of the whole brick.

Fig. 1.

AMERICAN CERAMIC SOCIETY.

34*

l Fig. 1 is shown a typical failure of a silica brick under the load

St.

A very interesting curve is obtained from a continuous record

the expansion and contraction of a silica brick during the load

st. In Fig. 2 is shown the theoretical expansion curve as given

r J. Spotts McDowell1 and the actual expansion curve as de-

rmined in a load test. A curve showing the behavior of a clay
'e brick under load is also given. The data from which curves
and 3 were plotted was not definitely determined but the general
end of the curves is unquestioned.

6. Crushing Strength.— The crushing strength (cold) of a
lica brick (on end) has often been used as a measure of the
uality of the bond. The crushing strength of the different
rands varies from 900 lbs. per sq. in. as a minimum to 3700
»s. per sq. in. as a maximum, the average being from 1800 to
too lbs. per sq. in.

1 Bull. Am. Inst. Mining Eng., p. 2044, November 1916 (No. 119).

$42 JOURNAL OF THE

7. Specific Gravity. I). W. Ross1 has proven the value of
the specific-gravity determination in the detection of under burneq
silica brick. The method for making this determination is the
one used for the apparent sp. gr. of ceramic materials the ap-
parent sp. gr. of a silica brick being quite close to its true sp. gr.
Using a 20-gram sample and boiling for one hour to obtain satura-
tion has been found to be satisfactory and as accurate as the two-
inch cube recommended by Mr. Ross. Check determinations
should be made to guard against error.

The sp.-gr. determination undoubtedly detects under-burning.
Samples taken from 14 separate cars of silica brick, known to
have been burned from cone 15 to 16, gave an average sp. gr. of
2.42, while samples from 12 separate cars of brick known to
have been burned to cone 18 gave 2.36. Four samples of brick,
thought to be under-burned, were re-burned to cone 20 and
tested. The results were as follows:

No.

Before re-burn
Sp. gr.

ing.

After

re-burning.
Sp. gr.

I

2.44

2.38

2

2 .40

2.36

3

2-39

2.36

4

2-43

2.36

8. Porosity. — The variation in the porosities of burned silica
brick is slight and little use can be made of this determination.
The porosity of most silica brick varies from 22 to 29 per cent.

Quite a number of other tests such as slagging, spalling, im-
pact, etc., have been used but the authors have not enough in-
formation to warrant their detailed discussion.

.

Effect of Impurities.

Lime. — The lime content of silica brick is well defined, abou
2 per cent, being necessary. If the lime content is less than 2
per cent, the bond will be weak and, if higher, the refractcriness
of the brick will be lowered.

Alumina. — Alumina, introduced in the form of clay or other
silicates, is probably the most active of all the fluxing impurities
present in a silica brick. It should not exceed 1.50 per cent. Its

1 Trans. Am. Cerant. Soc, 19, 83 (1917).

AMERICAN CERAMIC SOCIETY. 343

presence in excess of this amount is easily detected by a softening
point below cone 32. Although a low softening point is not proof
of an excessive alumina content — it should be the first impurity
looked for.

Iron. — Ordinary iron discoloration seems to have less effect
on the refractoriness of a silica brick than would be expected.
The iron present seems to become saturated with silica and remain
relatively inactive. The actual color depends considerably upon
the degree of oxidation. A chemical analysis of a dark brcwn
section of a silica brick gave 4.83 per cent. Fe203. This sample
was heated to 2800 ° F. and held at this temperature for one hour.
No spreading or local fusion was in evidence.

Two small chips from a brick showing iron spots were heated
to cone 32 started or just below the softening point of the brick.
The iron color seemed to distribute slightly but increased fusion
or slagging was not observed. A badly discolored chip
from a brick softened at about cone 32 or about half a
cone below the normal softening temperature. Although the
content of iron should not, in most cases, exceed 1.5 per cent., the
importance of the presence of this impurity in silica brick has been
exaggerated — due to its high coloring power.

The other impurities which are found present in very small
amounts in silica brick are usually not considered.

Safe Operating Temperatures for Silica Brick.

There is a tendency to claim a very high operating temperature
for silica brick, and statements are often made that a silica brick
will not soften at 31000 or 32000 F. Statements of this kind
should be qualified somewhat since in actual practice the bricks
are subjected to the higher temperatures on one side only, the
other side being much cooler. Owing to its high conductivity, the
whole brick does not reach the temperature of the furnace — the
heat being conducted away rapidly.

In order to obtain some information on this subject the fol-
lowing tests were made: A 3" by 3" by 9" section from each of
two shape bricks, manufactured by two of the largest companies
in the country, were heated to 2900 ° F. and held at this tern-

34 I

Ji IURNAL OF THE

perature for 24 hrs. The temperature during the run did not
vary more than 25 ° F. above or below 29000 F. Both bricks
softened and slumped down to the bottom of the furnace. In
Fig. 3 is shown the condition of these bricks after the test.

Fig. 3.

The test was repeated at 28000 F. and the bricks showed no
fusion, the only effect being a whitening of the surfaces. The
bricks were tested in a laboratory furnace and were exposed to
the furnace temperatures from all sides — the entire mass thus
attaining a practically uniform temperature. This indicates that
silica brick may be subjected to higher temperatures when laid
in a furnace wall in actual practice but the factor of safety lies
only in the fact that the bricks do not attain this temperature
throughout.

Laboratories of the H. Koppers Co.
and Pittsburgh By-Product Coke Co.

AMERICAN CERAMIC SOCIETY. 345

COMMUNICATED DISCUSSIONS.

A. A. Klein : The authors are correct in stating that the proper-
ties of the raw silica dominate the properties of the silica brick.
The complex thermal behavior of the quartz, including the modi-
fications in which the silica is usually present in the raw ma-
terial and the modifications of the silica resulting from the heat
treatment, account for the peculiar properties of silica brick, for
example — permanent expansion, spalling, etc.

The chemical analysis of a burned brick is of little assistance in
a determination of the value of the brick — since it does not dis-
tinguish between the percentages of the various modifications of
silica present. The importance of this point was shown forcibly
in an investigation by the writer of the constituents of various
standard brands of silica brick. The chemical analyses of the
brands examined were found to show no great variations but the
mineral constituents varied from, for instance, about 20 per
cent, of unchanged quartz remaining in the best brick to about 80
per cent, quartz in the poorest one. In some cases, large differ-
ences in constituents were found where not only were the analyses
similar but also where the same ganister formation served as the
source of the raw material.

The importance of determining the constituents is evident
from the great difference in the densities of the three silica modi-
fications. These have been accurately determined by many in-
vestigators and there is no question as to the large expansion
taking place when quartz inverts to cristobalite. The practical
application of this is evident. If bricks contain a considerable per-
centage of unchanged quartz after being burned, and are used
under conditions where the temperature reaches 1250° C. or
higher, the inversion of the quartz in the bricks continues to a
greater or lesser degree — depending upon the temperature. The
expansion of the bricks keeps pace with this inversion until a
dangerous condition of strain occurs — often resulting in great
material damage.

It is therefore essential that a direct or indirect quantitative
determination of the constituents be made. The easiest way
is to obtain the apparent specific gravity of the brick — which

346 JOURNAL OF THE

gives an accurate idea of the amount of unchanged quartz re

niaining and Mills measures how much expansion has been "burned
out" of the brick. This test is a most important one.

Although the chemical analysis of a burned brick is of little
importance, the fact remains that it may be used, with reserva-
tions, to determine whether a raw material is suitable for silica
brick-making. The authors speak of impurities causing a lower-
ing of the refractoriness. It must be emphasized here that a
silicious raw material may also be too pure. Perhaps an analogy,
drawn outside the field of silica brick, would be a comparison be-
tween the Austrian and California magnesites used in the manu-
facture of magnesite brick.

To illustrate the effect of over purity, in a case the writer has
in mind a certain quartzite containing 98 per cent, silica was
used in making silica brick. It was found that, although the burn-
ing had caused sufficient quartz inversion, the bricks were "punk"
and weak — due to the fact that there was not sufficient flux
present to completely bond the silica grains.

It is safe to state that a rock suitable for silica brick manufac-
ture should contain approximately 95 per cent. Si02, but it is
well to caution against the thought that any rock having this
silica content is satisfactory for this purpose. Other factors must
be taken into consideration — for example — the nature of the im-
purities, the structure of the rock, etc.

In conclusion, the writer is of the opinion that the statement
"under-burning may be roughly checked by linear measurements
and tapping with a hammer," made by the authors, under the
heading "Visual Inspection," should be qualified. Measuring
the brick will give rough checks only with bricks made at the same
plant and compared to the same standard. It has been the
writer's experience that very little reliance can be placed on the
"ring" of a silica brick as indicating under-burning — since well-
burned and under-burned bricks may have the same ring or the
same lack of ring.

R. M. Howe: Much valuable information is given by the
authors of this paper — their points concerning the inspection of
the raw material and the finished product being particularly

AMERICAN CERAMIC SOCIETY. 347

important. One man at a plant, who examines a bit of ques-
tionable raw material, who sees the cones fall during burning,
who observes the care exercised during the process and how care-
fully the bricks are sorted, can tell more about the quality of
the silica brick than a good-sized laboratory staff. His work is
largely that of control and, knowing exactly what he seeks, his
services are invaluable to the consumer. This is especially true
when his observations are supplemented by a laboratory.

The points made by the authors in regard to the chemical
analyses are well taken — as one cannot interpret the analytical
results more than approximately. A silica content of 96.0 per
cent., however, is rather unusual and should not be considered
a necessity.

A few typical analyses of silica brick of acknowledged quality
and which are the average of a large number of determinations,
are given as follows:

1. 2. 3. 4. 5.

Silica 96.05 95.85 94-96 95 58 95-31

Alumina 0.88 1.10 1.19 0.85 1.34

Iron oxide 0.79 1.04 0.85 1.26 1 .31

Lime 1.80 1.80 2.02 1.77 2.04

Magnesia 0.14 0.20 0.22 0.25 0.11

Alkalies 0.39 0.15 0.38 0.13 0.27

From the above analyses it is apparent that many high grade
ganister bricks contain less than 95 . o per cent, of silica. Analysis
No. 1 is of a particularly pure brand of silica brick. An aver-
age silica content of 94.5 per cent, is hardly too low for a good
grade of silica brick.

Softening points and load tests made on silica brick are of little
value. The softening points do not vary more than 2 cones and
the brick generally support a load of 25 pounds per square inch
at 15000 C. Any which do not may be eliminated by other
tests.

The crushing, strength test on silica brick gives a large amount
of information and can be verified by the modulus of rupture
test — the latter not requiring such a large testing machine.
These tests distinguish between well and poorly made brick, soft
and hard burned brick, good and poor burning silica-bearing ma-

348 JOURNAL OF THE

(crial (some form a punky structure when burned), and proper
and improper Lime content, etc. They also serve to detect, quali-
tatively, the presence of any clay and other fluxes which may pro-
duce a firmer bond than ordinarily encountered.

The specific-gravity test must be conducted with care. It is
generally sufficient to boil the test pieces in water at atmospheric
pressure. However, the following results show that ordinary
boiling is sufficient in one case and not in another: Each figure
represents the average of four determinations made on samples
taken from the same brick (the weight of the samples varying
from 20 to 30 grams).

Boiling
Boiling V'2 hr. Boiling 1 hi. Boiling 3 hrs. 3 hrs. and

and standing 16 at atmo- at atmo- then 1 hr.

Brand. hrs. in water, spheric pressure, spheric pressure. in vacuum.

Sample No. i. ... 2.212 sp.gr. 2.357 sp. gr. 2.367 sp. gr. 2.333 sp. gr.

Sample No. 2. .. . 2.165 sp. gr. 2-312sp.gr. 2.322 sp. gr. 2.320 sp. gr.

The following determinations also show that boiling under a
vacuum for 45 minutes is about equivalent to boiling 3 hours at
atmospheric pressure and then one hour in a vacuum:

Boiling 3 hrs.
at atmospheric
pressure and 1 hr. % hr. boiling

in a vacuum in a vacuum

Sample No. i 2 . 333 sp. gr. 2 . 356 sp. gr.

Sample No. 2 2 .320 sp. gr. 2.316 sp. gr.

Average 2 .327 2 .336 sp. gr.

In the cases cited, there is evidence that sample No. i also
needed a more thorough boiling. It was also the one which was
not saturated by boiling 3 hours in air at atmospheric pressure.

It is surprising to note that a re-heating test is not mentioned
by the authors. The specific-gravity test is of particular value
in indicating the relative conversion of quartzite into cristobalite
and tridymite and, incidentally, how well the expansion has been
"burned out" of the brick. This expansion is really the important
factor and can be determined directly by re-heating a sample for
five hours at 14000 C.

The authors evidently distinguish between general tests and
control tests. General tests include the chemical analysis, visual

AMERICAN CERAMIC SOCIETY. 349

inspection, load test, softening point test, specific-gravity de-
termination, re-heating test, crushing strength and modulus of
rupture tests. Many of these, including the chemical analysis,
softening point test, load test and others of a like nature, need not
be repeated unless in exceptional cases. However, the specific-
gravity test, checked by the expansion test, will always indicate
the conditions of the burn. The modulus of rupture will indicate
the strength of the material. These last three tests, when used
for purposes of control, are deemed sufficient for that purpose.

It is very interesting to note the temperatures at which the
silica brick tested by the authors softened. In reporting the re-
sults one factor has been overlooked. The fact that a silica
brick, when heated throughout, softens at 29000 F. does not
mean that it cannot be heated in use at 3000 ° F. or 3100 ° F. The
factor of heat conductivity has been overlooked. We have
softened cone 33 in a furnace made of material softening at cone
32. This furnace was thin walled and its conductivity was high.
Such a furnace, made from cone 32 material, gives good service
at cone 32 and cone ^^. A similar case is found in the malleable
furnace and many others in which the silica bricks are often sub-
jected to temperatures higher than their softening points. (Re-
vised schedule of cone softening temperatures.)

This state of affairs is more true of silica brick than of fire-clay
brick. At very high temperatures, the conductivity of silica
brick is such that, although they may soften at 29000 F. when
heated throughout, they give good service above 3000 ° F. if
not over-insulated. A thin "glaze" forms on the hot portion
of the brick but extends only for a short distance into the brick.
The portions not immediately in contact with the hot zone are
so protected that their softening temperature is not even ap-
proached, while the hot portion is so very viscous that it is not
deformed.

Mellon Institute of
Industrial Research.

D. W. Ross: The authors of this paper have indicated the
importance of familiarity with the raw material from which a
silica brick is made — when determining the nature of the brick

350 JOURNAL OF THE

and its value for any given purpose. This point is worthy of
serious attention.

The nature of the work which these authors have been doing
lias involved the testing of large numbers of siliea brick and in
this connection they have applied the specific-gravity test, nun
tioned in their paper, to brick made from quite a variety of raw
materials. It is gratifying to the writer that they report favora-
bly on this test. However, I am inclined to place more confi-
dence in the results of specific-gravity tests in which the test
pieces have been subjected to a vacuum (equivalent to 24" of mer-
cury) while immersed in hot water rather than in those in which
the test pieces have been merely boiled in water. It would also
appear advisable to use test pieces of a size such that determina-
tions upon them will represent the average properties of the brick,
thus avoiding the possibility of merely obtaining the properties
of a few abnormal particles.

Before recommending or condemning a brick on its softening
temperature, it would appear necessary to know the nature of
the raw material.

In a determination of the bonding strength, the cross breaking
test on silica brick placed on edge (6" between supports) appears
preferable to the compression test. Silica brick apparently fail
in this test through a tearing apart of the material below the
neutral axis rather than by the crushing of that above it. If,
perchance, a brick fails by shearing over one of the knife edges,
this can easily be detected and the results on that particular
specimen may be eliminated.

R. J. Montgomery: Mr. Klein's statement, as to establish-
ing the expansion of the brick upon burning, is correct and is re-
ferred to by the authors in the statement "under-burning may be
detected by linear measurements and tapping with a hammer."
The quality of any silica brick that does not have a good ring
is open to question and this is a good plant test — -even though it
does not always detect under-burned brick. A brick having a
poor ring is apt to be under-burned of lacking in bond.

Mr. Howe and Mr. Ross question the method used in making

AMERICAN CERAMIC SOCIETY. 351

le specific-gravity determination. M. F. Beecher1 reports
msiderable work on the various methods of obtaining satura-
on and concludes that "Saturation by boiling in water for 45
inutes to one hour is sufficient for the absorption determina-
on," when tests are being made on the same kind of materials
ir comparison, and that the "treatment with a vacuum does not
fer better results." The above confirms the experience of the
ithors. As to the size of the test pieces, used in making the
>ecific-gravity tests, correct results may be secured from 20-gram
imples providing that the results on two samples, taken from
fferent parts of a brick, check each other.

Specific-gravity determinations on four brick, from different
!ants and varying quite widely in the degree of burning, gave
ie following results:

Sample. Trial No. 2" cube. 1J<£" cube. %" cube. %" cube.

vSp. gr. Sp. gr. Sp. gr. Sp. gr.

2328 I 2.43 2.43 2.44 2.41

2 2 . 40 2 . 43 2 . 44 2.42

2150 I 2.27 2.29 2.3O 2.28

2 2.27 2 .28 2 .30 2 .28

1995 I 2.47 2.48 2.51 2.48

2 2 . 49 2 . 49 2 . 49 2 . 48

1997 I 2.31 2.33 2.34 2.32

2 2.3O 2.33 2.33 2.32

In making the above tests, cubes varying from 2" to V2" were
iken from each brick and subjected to the following boiling
eatment :

2" cubes weighing about 260 g. were boiled 4 hours
i3/t" cubes weighing about 65 g. were boiled 3 hours
3/t" cubes weighing about 18 g. were boiled 2 hours
V2" cubes weighing about 10 g. were boiled 1 hour

The above results indicate that if the specific-gravity determina-
011 is made on a 2" cube, boiling will not give complete satura-
on — as the higher values were obtained with the smaller cubes.

piece weighing from 1 8 to 65 grams gave the best results. The
isults on a 10-gram piece are not as reliable as those secured on
; Trans. Am. Ceram. Soc, 18, p. 73.

352 AMERICAN CERAMIC SOCIETY.

the medium sized pieces but check more closely than <1<> thai

on the 260-gram pieces.

Mr. Howe's reference to flame temperatures, as high as $000
F., reached in furnaces constructed of silica brick, brings out ai
important truth which has lead to the steatment by many manu
facturers and users that a silica brick will actually withstand tern
peratures above its softening point.

The re-burning test, noted by Mr. Howe, is referred to ii
the paper under the paragraph on specific gravity — some result
obtained by this test being tabulated.

3LYCHR0ME DECORATION OF TERRA COTTA WITH
SOLUBLE METALLIC SALTS.

By Hewitt Wilson Columbus, Ohio.

The terra cotta manufacturer is looking for some cheap, quick
id easy way of applying two or more colors to the same piece
ware. He does not have to be as particular as the potter in
e nicety of the lines and the delicacy of the colors. Most of
s work is placed beyond the range of close inspection and his
>lored areas are comparatively large.

The method of polychroming best suited to any given order of
rra cotta must be determined from (a) the size of the colored
eas, (b) the number of colors or colored areas on the same piece,
) the shape and finish of the outline of the colored areas, and
) the respective colors to be applied. Often a combination of
veral methods is necessary before the order is completed.
The present methods of applying colors are :

(i) Painting Colored Glazes onto the Required Areas with a
*ush. — This method is slow, expensive, and is liable to show
reaking due to the variation in thickness of the glaze. A great
any glazes of the Bristol type or variations of the same cannot
: painted without developing cracks on drying caused by the
levenness of the coating — these cracks do not heal in the kiln.
his method finds the greatest usefulness with glazes of low vis-
sity — which flow and cover well and which are usually burned
the lower temperatures. It is more adapted to small areas
id small ornamental features than to large ones.

(2) Spraying the Colored Glazes with Masques. — By this
ethod a paper, plaster or metal covering is placed on the part
)t to be coated at that time and the exposed portion is sprayed,
he placing and shifting of the masques injures the freshly sprayed
irfaces and leaves uneven and broken lines between adjoining
ilors. This is a good method for large polychrome surfaces

354 JOURNAL OF THE

and features in which the colors are arranged in simple repeating

design.

(3) Overglaze Spraying. In the first two methods, colored
glazes arc placed in their respective positions. By the third method
a single glaze may be sprayed over the whole piece and the over]
glaze colorant sprayed with a fine spray onto the desired par]
and brushed off the other parts. It is difficult, without the us]
of shields, to apply more than one color in this manner to the
same piece. It works well with small, complicated designs and
over large areas.1

(4) Painting with Soluble Metallic Salts. — This method
was probably adopted from the underglaze decorative process-
used for pottery and fine ware. In this process, soluble metallic
salts are dissolved in glycerine and painted onto the biscuited
or dry piece, dried to 1200 C in order to set the glycerine, and
fired to 800 ° C. The glaze is then applied and the piece burned
to the proper maturing temperature.

Purdy2 gives the following formulas for soluble underglaze
colors :

Grams.

Turquoise: Cupric nitrate 50

Glycerine 200

Water 80

Light blue: Cobalt nitrate 65

Glycerine 115

Water 50

Yellow: Uranic nitrate 65

Glycerine 60

Water 40

Yellowish brown: Ferric chloride 75

Glycerine 75

Water 50

Chamois: Nickel nitrate 100

Glycerine 100

Water 100

1 Trans. Am. Ceram. Soc, 19, 653 (1917).

2 "Class Notes" (1912).

AMERICAN CERAMIC SOCIETY. 355

Grams.

Gray: Cobaltic nitrate 3°

Ferric chloride 10

Glycerine 160

Water 4°

Green : Chromic acid i oo

Water ioo

For coloring terra cotta by this method, the piece is sprayed
th a single glaze and the various solutions are painted onto
J required areas. As the liquid quickly soaks into the dry
ize, no heating is required to set the glycerine. The ware is
rned in the usual way in a single fire.

Mediums. — When water is used, save with cobalt sulphate
d chromium acetate, the brush marks are too pronounced
;er firing. These two colorants spread easily and give uniform
lors.

Glycerine is not absorbed by the dry glaze as quickly as water
d allows the successive brushings to mingle. Alcohol may be
id for thinning the solution. Many of the salt solutions are
king in color so that the operator cannot detect the ground that
s been covered. Common bluing or soluble dyes cause the
inted areas to show distinctly.

Experimental.

Preparation. — The salts were weighed and ground together
a mortar. The required amounts of glycerine and alcohol
re added and the whole ground until the salts had disssolved.
te color was then applied on the dry glaze. If allowed to stand
the open, the concentration of the solutions is increased by
aporation and more coloring material will be added to the
ne area of base glaze with the same brush stroke. A less vola-
i medium than alcohol would be preferable for this reason.
any of the salts, for example some of the chlorides and sul-
ates, attack the carbonates in the glaze and cause bubbles
d blisters. The nitrates and acetates are usually non-active
this respect.

356

JOURNAL OF THE

Firing of Glazes and Colors. After painting, the decoran
trials were Set, with no special precautions, in ail ordinary ten
cotta kiln and fired with the regular ware to cone 6 7. Tl
glazes and underslip used were as follows:

Glazes.

Glaze "A" 0.268 K,0

0.468 CaO
0.264 ZnO

Glaze "B" 0.286 K20

0.403 CaO
o. 146 ZnO
0.074 MgO
0.091 BaO

Glaze "C" 0.238 K2O

0.354 CaO
0.210 ZnO
0.055 MgO
o. 143 BaO

0.417 AI2O3 \ 2.138 Si02

o . 446 AI2O3

o . 430 AI2O3

2.331 Si02
o. 104 SnOa

2 . 902 Si02
o. 141 S11O2

Vitreous Sup.1

Standard "K".

Flint 194

Cornwall-stone 437

China clay 242

Ball clay 117

990
"A" Series.

I. 2. 3. 4. 5. 6. 7. 8. 9.

Cobalt sulphate i.o 2.0 5.0 5.0 7.5 10.0 15.0 20.0 25.0 gran

Water 100. o 100. o 100. o 50.0 50.0 50.0 50.0 50.0 50.0 cc.

Glaze "A". — The colors developed varied from light blue
dark blue. Very even colors can be produced by the use
water alone although glycerine aids in the ease of painting ai
insures uniformity.

Glaze "C". — This produced grayer blues than those of Gla
"A", although the range of colors is the same.

AMERICAN CERAMIC SOCIETY. 357

"B" Series.

This was the same as the "A" series save that bluing was
;ed with the water in order to distinguish the painted areas
om the unpainted. The burned colors were the same.

"C" Series
1. 2. 3. 4. 5. 6. 7. 8. 9.

)balt sulphate 1.0 2.0 5.0 5.0 7.5 10. o 15.0 20.0 25.0 Grams

iromium acetate x.o 1.0 1 .0 0.5 0.5 0.5 0.5 0.5 0.5 Grams

ater 100. o 100. o 100. o 50.0 50.0 50.0 50.0 50.0 50.0 Cc.

The chromium solution was too weak to change the colors from
lose of the "A" and 'B'' series and the light blue of (1) was sim-
y deepened into the dark blues of 7, 8 and 9. A 30.0 per cent.
)balt solution (No. 7) gives the shade ordinarily desired.

"D" Series.
1. 2. 3. 4. 5. 6. 7. 8. 9.

obalt sulphate 10. o 9.0 8.0 7.0 6.0 5.0 4.0 3.0 0.0 Grams

ranium sulphate 0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 Grams

:atcr 50.0 50.0 50.0 50.0 50.0 50.0 50.0 50.0 50.0 Cc.

Glaze "A." — (1-5) Bright blues with green streaks where the
aze and color are heavy. (6-8) Greener but not uniform. (9)
le uranium yellow-green which develops well with this high zinc
aze. Uranium alone gives variable results when used on a
rge scale.

Glaze "B." — (1-3) Bright blue. (4-8) Good greens ranging
om blue-green to bright green. (9) Good uranium yellow
■een.

Glaze "C." — (1-6) Blues, gray-blues, blue-grays, (7-8) grays,
lottled yellow-brown (9).

Standard "K." — Greenish tinged blue or black-blue through
lue-gray to gray in (9).

"K" Series

1. 2. 3. 4. 5. 6. 7. 8. 9.

obalt sulphate 5.0 4.0 3.0 2.0 1.0 1.0 1.0 1.0 1.0 Grams

ranium nitrate 6.0 6.0 6.0 6.0 6.0 7.0 8.0 9.0 10. o Grams

lcohol 10. o 10. o 10. o 10. o 10. o 10. o 10. o 10. o 10. o Cc.

lycerine 10. o 10. o 10. o 10. o 10. o 10. o 10. o 10. o 10. o Cc.

358

JOURNAL OF THB

Glaze "A." -(1,2,3) Blue-greens. (4,5) Good greens. (6,7,8)
Yellow-greens. (9) Brown-green.

('.laze "B." — (1,2,3) Best greens. The remainder of the series
are yellow and yellow -brown-greens. All are good.

Glaze "C." — (1,2) Gray-blue. (3,4) Green-gray. (5,6) Green-
brown. (7,8,9) Dark green-brown.

vStandard Slip "K." — Blues to gray-browns. (5,6,7) Have a
slight green tinge. (3) Was flashed and a blue-black.

"F" Series.

1. 2. 3. 4. 5.

Antimony chloride io.o io.o io.o io.o io.o Cc.

Rntilc 0.0 0.5 1.0 1.5 2.0 Grams

Glycerine io.o io.o io.o io.o io.o Cc.

Soda ash 6.0 6.0 6.0 6.0 6.0 Grams

Glaze "A." — (1,2) Gray. (3,4,5) Yellow scummed.

Glaze "B." — (1) No color, scummed surface. (2) Faint orange
tinge. (3) Spotted orange. (4) Orange but variable. (5)
Crawled, variable rutile orange.

Glaze "C." — (1,2) Light gray. (3,4,5) Best yellows of all
glazes. The acid in the antimony chloride attacks the car-
bonates of the glazes and causes blistering. Hence the soda
ash was added to neutralize the acid without success.

Standard "K." — Yellow with good increase of orange colors
with the series. Blistering.

"G" Series.

1. 2. 3. 4. 5.

Cobalt sulphate 1.0 1.0 1.0 1.0 1.0 Grams

Uranium nitrate 6.0 6.0 6.0 6.0 6.0 Grams

Iron chloride 0.0 1.0 2.0 4.0 8.0 Grams

Alcohol 10.0 10.0 10.0 10.0 10.0 Cc.

Glycerine 10.0 10.0 10.0 10.0 10.0 Cc.

Glaze "A." — (1,2) Good greens. (3,4,5) Give lighter anc
browner colors.

Glaze "B." — (1) Dark almost black-green. (5) Yellow-brown
green. Good even colors.

Glaze "C." — Dark gray-greens to greenish browns. The iroi
affords an individual brown color in spots.

AMERICAN CERAMIC SOCIETY. 359

Standard "K." — (1) Gray-black to light neutral gray. (5) Iron
cts as a diluent and does not intensify the color.

"H" Series.

1. 2. 3. 4. 5.

Cobalt sulphate 1.0 1.0 1.0 1.0 1.0 Grams

Jranium nitrate 6.0 6.0 6.0 6.0 6.0 Grams

ron chloride 2.0 2.0 2.0 2.0 2.0 Grams

Chromium acetate 0.0 0.5 1.0 2.0 4.0 Grams

ilcohol 10.0 150 150 150 15.0 Cc.

Glycerine. . . : 10.0 10.0 10.0 10.0 io.oCc.

Glaze "A." — -(1,2) Uranium greens. (3,4,5) Green-browns.
Glaze "B." — -(1,2) Yellow-green. (3,4) Yellow-brown. (5)
5uff- brown.

Glaze "C." — (1) Gray-green. (2) Brown-green. (3,4,5)
browns.

Standard "K." — (1) Dark blue-gray. (2,3,4,5) Grays and
•lacks.

"I" Series.

1. 2. 3. 4 5.

Jranium nitrate 4.0 2.0 1.0 0.5 0.0 Grams

ron chloride 0.0 0.5 1.0 2.0 4.0 Grams

dcohol 10.0 10.0 10.0 10.0 10.0 Cc.

Mycerine 10.0 10. o 10.0 10.0 10.0 Cc.

Glaze "A." — (1,2) Varying uranium greens. (3) Light yellow.
4,5) Light buffs.

Glaze "B." — (1,2) More orange than Glaze "A." (3) Bright
-ellow. (4,5) Light buffs.

Glaze "C." — -(1,2) Yellow and brown-green. (3) Light buff-
rellow. (4,5) Brown edged yellow-buffs.

Standard "K." — {1,2) Blue-grays. (3,4) Light buffs. (5)
)ood tan.

"J" Series.

1. 2. 3. 4. 5.

fin chloride 5 .0 10.0 15 .0 20.0 25 .0 Grams

Chromium acetate saturated in alcohol 0.0 5.0 5.0 5.0 5.0 Cc.

Mycerine 10.0 10.0 10.0 10.0 10.0 Cc.

Ucohol 5.0 7.5 15.0 20.0 25 . o Cc

36o JOURNAL OF THE

Glazes "B" and "C." Pink colors but crow tooled, dry and
blistered. Some of the blistering occurs in the application.
"K" scries is good with (4) of "J" series.

"K" Series.

1. 2. 3 4. 5.

Tin chloride 20.0 20.0 20.0 20.0 20.0 Grams

Chromium acetate saturated in alcohol 0.0 2.0 4.0 8.0 16.0 Cc.

Glycerine 10.0 10.0 10. o 10. o 10.0 Cc.

Alcohol 25 .0 23 .0 21 .0 17 .0 9.0 Cc.

Glaze "B." — (1) Light pink but crowfooted. (2-5) Increasing
pinkness. Very good colors.

Glaze "C." — -(1,2) Light pinks, dry glaze. (3,4,5) Increasing
pinkness but with a buff tinge.

"L," Series.

1. 2. 3. 4. 5.

Tin chloride '. 20.0 20.0 20.0 20.0 20.0 Grams

Chromium acetate (5 per cent, in H20) . 2.0 4.0 8.0 16.0 25.0 Cc.

Alcohol 8.0 6.0 2.0 0.0 0.0 Cc.

Glycerine 10. o 10.0 10. o 10. o 10.0 Cc.

Glaze "B." — -Pink colors are good and equal to those of "K"
series.

Glaze "C." — (1,2) Light pink, liable to crowfoot, blister and be
dry. (3,4,5) Very light pink, lighter than "K" series.

"M" Series.

1. 2. 3. 4. 5.

Manganese sulphate 5.0 10. o 15.0 20.0 25.0 Grams

Alcohol 5.0 10. o 10. o 10. o 10.0 Cc.

Glycerine 10.0 10. o 10.0 10. o io.oCc.

Standard "K." — (1) Light gray. (2) Dark brown glazed sur-
face. (3-5) Increasing dark brown variegated colors which can
be used if the running of the glaze is allowed for. Manganese
lowers the viscosity considerably.

Glazes "B" and "C." — -(i) Medium brown to streaked browns
and blacks in (3,4,5). Manganese sulphate melts into a glaze
by itself, whether in standard or glaze.

AMERICAN CERAMIC SOCIETY.

361

"N" Series.

1. 2. 3. 4. 5.

issium permanganate 5.0 10.0 15.0 20.0 25.0 Grams

hoi 50 10.0 10.0 16.0 25.0 Cc.

Brine 10.0 10.0 15.0 18.0 18.0 Cc.

hese mixtures become warm with emission of a white vapor
bubble when mixed with alcohol. This does not work as
. as the sulphate, showing blistering, peeling and dryness.
ors are not even and brush marks show.

"O" Series.

1. 2. 3. 4 5.

issium permanganate 0.0 0.5 1.0 1.5 2.0 Grams

lit sulphate 1.0 1.0 1.0 1.0 1.0 Grams

chloride 6.0 6.0 6.0 6.0 6.0 Grams

jmium acetate 0.5 1.0 1.5 2.0 2.5 Grams

:erine 10. o 10. o 12.0 15.0 15.0 Cc.

hoi 50 5.0 7.5 10. o 10.0 Cc.

rlaze "B." — Light and darker mottled brown-grays to a dark

vn.

rlaze "C." — Good red-browns, the color varying with the

iting.

tandard "K." — (1) Gray. (2,3,4,5) are dark browns and

ks.

"P" Series.

1. 2. 3. 4. 5.

lit sulphate 5.0 3.0 1.0 1.0 1.0 Grams

chloride 6.0 6.0 6.0 8.0 10.0 Grams

erine 10. o 10. o 10. o 10. o 10. o Cc.

hoi 10. o 10. o 10. o 10. o io.oCc.

laze "B." — (1,2) Violet-blues. (3) Light blue-gray. (4,5)
f, blue.

laze "C." — (1,2 ) Gray-blues. (3,4,5) Light blue-grays,
tandard "K." — (1) Flashed, blue-gray. (2) Dark green-black
led. (3,4,5) Good browns.

"Q" Series.

1. 2. 3. 4. 5.

le 0.2 0.5 1.0 2.0 4.0 Grams

erine 10. o 10. o 10. o 10. o io.oCc.

362 JOURNAL OF THE

Glaze "B." (i,2) Faint streaks of yellow. (3) Darker al

strcaki-d.

Glaze "C."— (4,5) Rutile yellow-brown, badly wrinkled.

Standard "K." — Good even colors ranging from light brow
to red-brown. May be used. As the rutile did not dissolv
this was an overglaze decoration rather than a soluble stain.

"R" Series.

1. 2. 3. 4. 5.

Iron chloride 2.0 4.0 8.0 16.0 25.0 Grams

Alcohol 10.0 10. o 10.0 10. o io.oCc.

Glycerine 10.0 10. o 10.0 10.0 io.oCc.

Glaze "B." — (1,2) Very light yellow. (3) Light brow
(4,5) Dark red-brown.

Glaze "C." — (1,2) Light buffs. (3,4,5) Increasing red-browm

Standard "K." — (1,2,3) Light gray-buffs. (4,5) Red-brown
varying.

All are good colors but vary as the iron volatilizes in burnin]

"S" Series.

1. 2. 3. 4. 5.

Copper sulphate 2.0 4.0 8.0 16.0 25.0 Grams

Alcohol 10. o 10. o 10. o 10. o 10.0 Cc.

Glycerine 10. o 10. o 10. o 10. o 10.0 Cc.

Glaze "B." — Colors duplicate those of glaze "C" but she
more black and not too much good green. This is just the r<
verse of the uranium series in which Glaze "C" gives the p<x
greens.

Glaze "C." — (1) Light copper-green, good and uniform. (:
Varying copper-green, fair. (3,4,5) Green-blacks and black. N
defects.

Standard "K." — (1,2) Just faintly off color. (3) Light greeni:
shades, blistered. (4,5) Uniform black.

"T" Series.

1. 2. 3. 4. 5.

Nickel sulphate 2.0 4.0 8.0 16.0 25.0 Grams

Alcohol 10. o 10. o 10. o 10. o 10.0 Cc.

Glycerine 10. o 10. o 10. o 10. o 10.0 Cc.

AMERICAN CERAMIC SOCIETY. 363

Glaze "B." — (1) Good nickel-purple, uniform. (2) Lighter
han (1). (3) Purplish gray. (4,5) Peeled and showing un-
lelted green dust which rubs off easily.

Glaze "C." — (1,3) Good browns, uniform. (4,5) Unmelted,
11b off.

Standard "K." — Good browns. (4,5) Greenish unmelted dust.

From the above series, the best colors were selected and used
1 painting ornamental terra cotta, including bits of scenery.

An unskilled man, not used to handling a brush, did very good
'ork after a few days' practice.

RESULTS.

Colors with the Glazes.

Blue. — Good blues are very easily produced with cobalt
ulphate in all of the glazes. Varying the amount of the sul-
fate from 0.5 g. to 30.0 g. per 100 cc. of water will give a
ange from a very light sky-blue to the ordinary deep blue. Be-
on d this amount the color darkens rapidly to blue-blacks.

Green. — Because of the zinc used in the glazes of this study,
hromium salts gave tans and browns which are stable, uniform,
nd easy to use. However, the green from chromium in zincless
lazes was not good and could not be relied upon.

The best green colors were produced from the uranium-cobalt
ombination in the mat glazes "A" and "B". The gloss glaze "C"
ave gray-greens. The best proportions were uranium nitrate 6
rams, cobalt sulphate 1 .0 to 3.0 grams, alcohol 10 cc. and gly-
erine 10 cc.

The copper-greens were variable and faded. However, Glaze
C", which gave poor greens with cobalt and uranium, gave better
esults with the copper.

Pink. — Uniform light pink colors can be produced in the above,
lazes with mixtures of tin chloride and chromium acetate. 20
rams of tin chloride, 16 cc. of a saturated solution of chromium
cetate in alcohol, with 10 cc. of glycerine and 9 cc. of alcohol,
;ave good light pinks.

364 J« >URNAL OF THE

Browns. Manganese sulphate to g., alcohol 10 cc. and
glycerine i<> cc. gave fair browns. If made stronger than this
or if painted heavily, it will darken from brown to black and run
when applied on an inclined surface. If more dilute the man
ganese will produce good grays. These browns, grays and blacks
may be varied by blending with cobalt, iron and chromium
in the "O" series.

Nickel Sulphate with Glazes "A" and "B" produced light purple
and gray triors1 but with (Maze "C" it gave good uniform browns.

Chromium Acetate in a glaze containing zinc gives good uni-
form tans and browns.

Iron Chloride gave red-browns but these were variable at the
firing temperature.

Yellow. — Dilute uranium sulphate or nitrate alone or with*
iron chloride gave promise of fair light yellow shades. However,
when tried out further they were not very reliable.

Gray. — This may be produced from several sources depend-
ing on the shade desired. Blue-grays were obtained in Glaze "C"
with cobalt and uranium — the uranium content being lower and
the cobalt higher than used for the production of green colors.
(See vSeries "D.")

Violet. — Cobalt and iron in "P" series gave violet and blue-
grays with a ratio of cobalt sulphate to iron chloride varying
from 5 : 6 to 1:10, respectively. Dilute manganese gives brown-
ish grays.

Colors with Vitreous Slip "K."

Blues are easily obtained with cobalt sulphate. Most of the
other salts or combinations give grays, browns or blacks. The
colors are not as bright as those developed in the glazes.

Browns. — Dilute uranium gave gray, strong uranium gave
gray-browns.

Uranium and cobalt combinations gave blue and green-blacks.

Iron gives buffs and red-browns.

Manganese gives gravs, browns and blacks, of different shades
however, than those produced in the glazes.
1 Trans. Am. Ceratn. Soc, 14, 141 (1912).

AMERICAN CERAMIC SOCIETY. 365

Rutile painted on the surface with glycerine gives good light
uffs and buff-browns.

Copper sulphate gives light green-grays running into blacks.

Nickel gave good browns.

In conclusion, we wish to state that the Denver Terra Cotta
ompany has also used this method on shop orders with success
id in fact is the pioneer for this method of color decoration of
'rra cotta on a large scale.

Ohio Statu University.

DISCUSSION.

Mr. Burt: I would like to ask Mr. Wilson if he has not ex-
;rienced large variations in color when using the glazes con-
lining the soluble salts. In other words — does not the per-
>nal factor influence the depth of color secured by spraying with
le brush?

Mr. Wilson: The amount of solution on the brush, the speed
' the brush across the absorbent surface of the dry glaze and the
:peated number of strokes over the same surface, all have an in-
lence on the amount of coloring material added to a given
mare inch of glaze surface.

If the concentration of the solution is so made that a moderate
>aking action of the glaze surface is necessary to produce the
;sired depth of color, more uniform results can be expected
Dth in regard to the shade and to the elimination of the brush
arkings. This aids in the solution of the personal factor prob-
m.

Mr. Truman: I should like to ask Mr. Wilson if, in weighing
at the soluble salts, he did not experience some difficulties through
leir absorption of moisture from the atmosphere5 I have ex-
erienced this difficulty in weighing out chloride of iron.

Mr. Wilson: Some of the salts absorb moisture very rapidly.
lhey must be weighed very quickly.

Mr. Truman: In some of my work in which soluble salts
rere used, I found it advantageous to add the salts in the form
f saturated solutions of predetermined specific gravity.

AMERICAN CERAMIC SOCIETY.

Mr. Mintox: I would ask Mr. Wilson whether he lias deter
mined the relative behavior of cobalt borate and cobalt sulphate

in his glaze work? I understand that the cobalt borate is to be
preferred.

Mr. Wilson: I have encountered some blistering of the glazes
when using chromium sulphate. I did not experience this diffi-
culty when using cobalt sulphate. I can see no objection to the
use of cobalt borate although I have not used it.

Mr. Radcliffe: My experience has been similar to that of
Mr. Truman's in regard to the variation in color, and I have
found that the best method was to use solutions of the soluble
salts, controlling their specific gravities by means of a hygrometer.
In this way the variation in color was reduced to a minimum.
Some of these soluble salts do cause blistering of the glazes dur-
ing their application — due to the acid in the salts. This difficulty
may be overcome by the addition of a metallic oxide to the solu-
tion and neutralizing the acid.

I

JOURNAL

OF THE

lMERICAN ceramic society

A monthly journal devoted to the arts and sciences related to
e silicate industries.

I. 1 June, 1918 No. 6

EDITORIALS.

:lay products at the chemical EXPOSITION.1

A survey of the exhibits at the Grand Central Palace revealed
| fact that while many products of the ceramic industries,
ler than glass, are indispensable in the processes of chemical
inufacture and research, the exhibits of these were not very
merous. There were found at the Exposition three displays
chemical porcelain, two of chemical stoneware, one of fire
ick, three of laboratory refractories, one of insulating material,
e of vitrified cells for acid towers, and two of enameled iron.

Chemical Porcelain.

The exhibits of chemical porcelain were those of the Coors
•rcelain Company, the Guernsey Earthenware Company, and

I ( )hio Pottery Company. Inasmuch as no criterion for chem-

II porcelain can be advanced without the evidence afforded by
ecial tests, it is not possible in a review such as this to make
y mention of quality. It will suffice, therefore, to describe
e three exhibits as one, especially as the same excellence in
inufacture and resource is evident in each case. The im-
Dvements effected since the Exposition of last year were quite
irked. All the factories are now successfully producing evap-

i Fourth National Exposition of Chemical Industries, Grand Central
lace, week of .September 23, 19 18. Reference to the Chemical Exposi-
I in the June number of This Journal is accounted for by the fact that the
uiary number of Vol. 1 o." the Journal was issued in July of this year — the
needing numbers being correspondingly late.

368 JOURNAL OF THE

orating dishes, both ilat and curved, of twenty-inch diameter
while of course the usual laboratory sizes of these, together with
beakers, crucibles, casseroles and spatulas, were strongly ir
evidence. The mastery of the technical difficulties in the work
shop is well shown by the excellent production of delicate porci
lain plates with minute perforations. Some of these form th<
bottoms of Gooch crucibles, while others are detached. Cer
tainly no finer specimens of workmanship are made anywhere.

Chemical Stoneware.

The two exhibits of chemical stoneware may likewise be coil'
sidered together. They were by the General Ceramics Com
pany, of Keasbey, New Jersey, and Maurice A. Knight, of Eas
Akron, Ohio. The modern chemical stoneware is thorough 1)
vitrified throughout. It is said that the glaze is added for ap
pearance only, but the smooth polished surfaces of pump plunger
and valve balls is satisfactory both to the sight and touch. Th<
structural intricacy of some of the pieces is a marvel to any ob
server, but to those who are versed in the especial difficultie;
of clay-shaping and firing, it is a veritable miracle. No problerr
of construction seems to baffle these potters; huge tanks and jar;
which would drown a goodly proportion of the Forty Thieves
condensing worms larger than any barrel, valves, spigots, traps
siphons in great variety, and all of them perfect, sonorous anc
acid-proof pottery. Certainly there is no need for importec
wares in this department of ceramics.

The exhibit of vitrified packing for acid towers seemed to b
something new. The B. Mifflin Hood Company, of Atlanta
Georgia, have produced, apparently from an auger die, a tub
about two inches in diameter which contains a spiral coil. Th
ingenuity of the product itself is quite fascinating and, given
burned clay perfectly resistant to the acid, the piece would seer
to be ideal for its purpose. A large exposed surface is combine
with a minimum weight.

Refractories.

The only exhibit of fire brick was that of the Didier-Marc
Company of Keasbey, N.J. The wares, so far as inspection migr

AMERICAN CERAMIC SOCIETY. 369

irrant, were of excellent quality and the only regret is that other
anufacturers did not exhibit. Fire brick are not, of course, an
cessory special to chemical manufacture and yet, in these days
high temperatures, chemists have a very real interest in the
lality of refractories.

Of more direct interest were the displays of laboratory refrac-
ries by the Carborundum Company, of Niagara Falls, New
3rk, the Denver Fire Clay Company, of Denver, Colorado, and
e Norton Company, of Worcester, Massachusetts. The first
,med showed especially the larger pieces, such as muffles and
ucibles of carborundum bonded with clay. The Denver Fire
ay Company had an excellent display of fire clay crucibles of all
:es, and a complete line of cupels and scorifiers. The Norton
>mpany specialized in the smaller wares, some of which were
tricate in form and all were well made — -crucibles and boats for
alytical work, tubes for electric furnaces and pyrometers, as
;11 as small cones and thimbles for special operations.
The Sil-O-Cel Company showed a collection of insulating ma-
rial with which an interesting and effective exhibit was made,
veral illustrative units were shown in which the insulating
operties of the material were compared with those of fire brick
d other standard materials.

THE GLASSWARE EXHIBITS AT THE EXPOSITION.

Among the ceramic exhibits, those of glassware proved to be
high interest. Since the exhibition was limited to prod-
ts of direct relation to the chemical industries, the glassware
hibits represented a very small field of the entire glass industry,
it this field is one of the greatest importance to the chemical
dustries and to those industries in which chemists and chemical
boratories are essential.

The glassware exhibits comprised three groups, namely, ex-
bits by glass manufacturers; exhibits by apparatus dealers, by
horn the bulk of the special apparatus is made before the lamp;
id exhibits by manufacturers of fused silica.
Of the glass manufacturers, the exhibit of the Corning Glass
Dmpany was the largest and most elaborate. A variety of
^yrex" laboratory apparatus and baking dishes formed the

370 JOURNAL OF THE

basis of the display. In addition, there were exhibits of "FrenselM
lenses and other signaling lights in a variety of colors as well as
of lenses for automobile and other headlights. ( )ne of the fea-
tures of this exhibit was the glass tubing for special purposes.
In this was included tubing for burettes and "lens" tubing e pi
cially for clinical thermometers. Demonstration was also made
of a lamp equipped with daylight glass. Popular attention was
attracted by a glass blower who demonstrated the simplicity
of working glass before the lamp.

The booth of the Macbeth-Evans Glass Company was del
voted to an exhibit of their well-known laboratory glassware.
The exhibit was simple, well chosen, and displayed the usual
excellent workmanship. One feature of the exhibit was a twentyl
inch "balloon flask" (18 gallons), used in mixing fulminate com-
positions.

( )ne exhibit of the Whitall Tatum Glass Co. consisted of a sim-
ple display of the long and favorably known "non sol" chemical
laboratory apparatus, together with a selection of reagent bot-
tles in a variety of forms.

Of the dealers' exhibits, that of the Emil Greiner Co. was of
commanding interest. It might almost be characterized by "any-
thing with a glass stopcock." This was not, however, the limit
of the exhibit. Thermometers, intricate absorption apparatus,
condensers, fractionating columns and Dewar flasks, were fea-
tures.

The glass exhibit of the Central Scientific Co. comprised only
a small portion of their total display. One exhibit was of inter-
est, however, in view of the standardization of sizes and shapes
of laboratory apparatus and the construction of special apparatus
in accordance with the specifications of American scientists.

An industry comparatively new to America is the production
of fused silica and quartz glass. The former, the opaque product,
and the latter, the transparent variety, were exhibited by the
Sidio Co. of America and by the Hanovia Chemical and Manu-
facturing Co., respectively. Although these exhibits were limited
to small- and medium-sized apparatus, the excellence of work
manship indicates that real progress is being made in this new anc
important industry.

AMERICAN CERAMIC SOCIETY. 371

The Thermal Syndicate also presented an exhibit of their well-
cnown produets. This exhibit was characterized by a display
)f the larger sized apparatus. The features were a nitric acid
•ondensing system, of elaborate construction, and a display of
Irawn tubing in ten-foot lengths and ranging up to four inches
n diameter. A line of the smaller articles in transparent silica
las recently been added.

Akin to the exhibits of glassware were those of enameled
ron apparatus displayed by the Elyria Enameled Products Co.
ind by the Pfaudler Co. These exhibits deserve more than the
tassing mention here given them, for they represent real progress
n an important field of ceramics.

The glass exhibits, as a whole, were pleasing and were credita-
)le to the American industries. It is to be regretted that a larger
lumber of manufacturers of glass apparatus were not represented,
mt the exhibits demonstrate that the necessary technical knowl-
:dge is not wanting in this line. It is reassuring to observe
hat the principal handicap, the lack of manual and mechanical
kill, is rapidly being overcome and that a permanent American
ndustry in scientific glassware is being surely established.
AMERICAN CERAMIC SOCIETY EXHIBIT AT THE
EXPOSITION.

By C. H. Kerr.

One of the features of the Ceramic Section of the Exposition
vas a series of eight charts displayed upon the walls of the Amer-
can Ceramic Society's booth. The interest and comment
vhich the charts aroused at the Exposition seemed to warrant
heir reproduction in the Journal and they are therefore presented
n this issue.

Very few, even of the members of the American Ceramic So-
•iety, have any clear conception of the great magnitude of the
ceramic industries, either considered as separate branches of the
ndustry, such as fire brick, pottery, glass, cement, etc., or con-
lidered in the aggregate, and it is not at all surprising, there-
ore, that the general chemist or the specialist in some other
branch of applied science should be quite startled to learn that
he value of ceramic products is about one-third greater than
die value of electrical machinery, about two and a half times

372 JOURNAL OF THE

the value <>i" the potato crop of the Nation, and more than twice
the value of the petroleum products of the country.

While the charts show the approximate figures, it may be of
interest to record the values, in round numbers, from which the
charts were prepared. The figures are the Government official
figures, taken from the Census Reports for the year 1914. Later
reports are available in many instances but the World War has
created such havoc with industrial statistics that the older figures
are considered more reliable in showing the relative proportions
of the various industries.

For the Manufactured Products the figures used in Chart No.
1 were:

Cotton goods $701 ,000,000

Men's clothing 584,000,000

Chemical industries 548,000,000

Ceramic industries 447 , 000 , 000

Petroleum refining 396,000,000

Woolen goods 395 ,000,000

Electrical machinery 335 ,000,000

Blast furnace products 318,000,000

For Mineral Products the figures used in Chart No. 2 were:

Coal, soft $493,000,000

Ceramic products 447,000,000

Iron ore . . 299,000,000

Petroleum 214,000,000

Coal, hard 188,000,000

Copper 153,000,000

Gold 94 , 000 , 000

Silver 40 , 000 , 000

The statistics regarding Agricultural Products (Chart No. 3)
are of the greatest interest, for, after all, the United States is
chiefly an agricultural nation though we are very apt to over-
look that fact because the agricultural production is made up
of innumerable small units rather than large, segregated groups.
We do not have a farm or a group of farms to compare with the
United States Steel Corporation, but we do have the grand

AMERICAN CERAMIC SOCIETY. 373

totals of agricultural products that dwarf even our immense
totals in manufactured products. The charts showing the rela-
tive value of Ceramic Products in comparison with agricultural
Products is based upon the following figures:

Corn $1 ,722 ,000,000

Wheat 878,000,000

Hay 779,000,000

Cotton 720,000,000

Oats 499 , 000 , 000

Ceramic products 447 ,000 ,000

Potatoes 199,000,000

Tobacco 96,000,000

In all of the above statistics there are included in the ceramic
industries the manufacture of clay wares of all kinds, glass, cement,
enameled wares, plasters, lime, abrasive products and allied
articles usually understood as being embraced within the ceramic
field. To analyze the items, making up the impressive aggre-
gate of 447 million dollars a year for the ceramic industries, would
be very instructive and very surprising, but that must be left
for a later issue of the Journal.

The most interesting curve shown at the Society's booth
was that which displayed in graphic form the growth of
the American Ceramic Society (Chart No. 4). Founded in
1898 and starting its real existence in 1899, the membership
shows a steady and consistent growth to somewhat over 500 in
1917. In 1918 the growth has been unusual and the member-
ship has more than doubled. At the time of the Exposition the
Society had about 1 100 members enrolled. It is eminently
fitting that the ceramic industry, one of the most ancient of all
the manufacturing arts, should have as a clearing, house for its
scientific activities, an organization with the size and power of
the American Ceramic Society.

Very few people, even among those engaged in ceramic work,
realize the great number of ceramic products which involve the
use of the various every-day raw materials, and still fewer ap-
preciate how exceedingly vital the ceramic industry is today in
supplying the products upon which other industries depend. These

374

JOURNAL OF THE

O ex

O

o
o

1 a. I— i-

a: ^2cc
d. Co.

O u-i

oc or

a: 5

Z -

O

U5

UJ Q

I
I-

o

-I

O

Z

WO

o

Z

<

tf>

r =

x

v3

Z

LU

cc

I-

UJ

O
O

— O UJ

ui O

UJ

Z

X

<

I

<

VJ

DC

I-

O

§

a

uj ^O,^

<
z2

§££

"" ^
<<^

o o
o

©
©

AMERICAN CERAMIC SOCIETY.

375

1-
o

UJ

5*S

i-

O

O
to

<

o

PRODUCTS
)0 50

4)

0DUC1
cts Com i

ODUCTS

CERAMIC
IRON ORE

LEU M

D

00 4<

a- <

^^^ ■■» ill

c
c
i-

L
C

r L. i r\ w
OAL , HAR
PPER

30 3

f dollars

Chart No. 2.

<|2

V.

J c

GOLD
LR

///0A75 0

Ml

e Value (
;ll Knov

5ILVE

0 l(

Ml

UJ

JOURNAL OF THE

u s

O O =3

a ~ o

a- ££

3 -J

° a
-I a «

<^£

J « I

—J hi vD
ID VJ <

2 3 I

«u ^

X

z
o

uj

££

I

Z

o

h
h
o

o

o

cc

Q.

I-
<

O

I
<.

cc

Ml

LU

o

<

I-
O

Q.

o
o
ao

o
o

2^

3*5:

O

.§■§ a

■^»

o ^
o

AMERICAN CERAMIC SOCIETY.

377

h

0
a.

O

I
«/>
cc

Ul
CO

I

UJ

ciety's
899

aO

o —

2
^

" 2i

<

a Z
111 O

o

5<

" 2}

3 y

a: Z

UJ ^

2>

3

VD O "

2

CM

o

ii

X io

o
o

Curve S
Growth

> c

> c

J c

> c

> c

> c

> €

5 M

! 5

> <

>
J

>

378

[OURNAL OF THE

v^^

AMERICAN CERAMIC SOCIETY

379

>-

D

z3

— O

O

i-2

38o

JOURNAL OF THE

LLl S *

-OH

arz*

v>k: 2

-* H O

O i- =

Zui 2.
ctt <

-as

UJ yj

Ul - o

ssgt

lllsft

%

I

~5s

<u

I

X

I

^^^^

S: vj s~ X Vj Q

v>

111

a
<

111 J
1*8 J

til

III £

I
1

4*

Vi

1
l

I
in

<

Si

IS

^ <

9

s

£* 1

-5^ Sc-

•5: ^

v> Z

5 z

-J Id

O %J

A ^ IL

<3hl Id

*** > cc

-J — J

III <S) <

z 5 a

■is

I

Ul V)

a.
S >-
-1 O

10

a 3
o 9

z

<

? Z ?: 2

v»

AMERICAN CERAMIC SOCIETY. 381

t .**

- £ -t §. 1 s -fell .^|^

> H

\ r ui ii (ft irt f n>U °

1 7 O i.i

382 J< )l k\Al, OF THE

two relationships are shown i:i two of the charts (Nos. 5 and 6)
which have been reproduced, l>ui it is obvious thai it lias not
been possible to include in such condensed form anything like
a complete list, either of the raw materials and their uses, 01 ol thd
other industries dependent upon ceramic products. There are,
however, many very interesting points shown. We are apt to toilet
that the manufacture of explosives is absolutely dependent upon
chemical stoneware and enameled iron with which to carry on
the manufacturing processes; that gas engines are useless with-
out spark plugs; that all metallurgical processes, especially oul
iron and steel industries, are dependent upon serviceable re-
fractories; that our ability to live comfortably in houses depends
upon glass for windows; that our entire city life is based on good
sanitation which is made possible only by the use of ceramic
products in the sewerage systems, etc. This dependency upon
the ceramic industries might be shown in almost innumerable
instances, but even the limited number shown in the charts is a
very convincing argument of the vital need of ceramic products
in our modern scheme of life.

The chart (No. 7) listing the better known branches of the
ceramic industry is self-explanatory, but the one (No. 8) which
tabulates a few of the vital War products calls for a little further
discussion. From what has been said in connection with the
other charts, it is quite plain that any list that pretends to give,
with any degree of completeness, the ceramic products vital to
the prosecution of the War, would necessarily be a very long and
complicated one, but in order to bring to public attention a few
of the most obvious or newest products the chart of Vital War
Products was prepared. Ordinarily, one would not think of
the refractories in the boiler room of any of our battle-craft as
being of more than passing import, but this problem has been
exceedingly critical and has now been satisfactorily solved.
For airplane engines, a new and better spark plug was required,
and it has been and is being produced. Optical glass has had
national attention and everyone is familiar with the story ol
the American manufacturer having successfully produced th(
required types, qualities and quantities, so that optical glass i;
not a limiting feature in the National War program. Labora

AMERICAN CERAMIC SOCIETY. 383

tory glassware and porcelain of American make are now avail-
able in satisfactory quantity and of good quality. And so the
list could be extended to the limit of space available, or to the
limit of the reader's patience, but the importance of these vital
products should be indelibly impressed upon the public at large
so that public opinion will see to it that in future Governmental
action these all-important industries are thoroughly appreciated
and their encouragement and continuance thereby guaranteed.

ORIGINAL PAPERS AND DISCUSSIONS.

THE POROSITY AND VOLUME CHANGES OF CLAY
FIRE BRICK AT FURNACE TEMPERATURES.1

By George A. Loomis.

I. Introduction.

General Considerations Regarding the Testing of Fire Bricks.—

The testing and classification of clay fire bricks is a subject which
has received considerable attention in recent years, but which has
not been worked out as yet — to any great degree of complete-
ness. Failure to satisfactorily classify clay fire bricks has been
due in part to the fact that clay refractories are used under such
a variety of conditions. A fire brick which proves very satis-
factory under certain conditions of use may not be able to fulfill
the requirements for other uses. Fire bricks may be called upon
to resist deformation under excessive or very light loads at high
temperatures, to resist sudden changes in temperature, to with-
stand the intrusion of slags or glasses, to resist abrasive action,
and many other equally severe conditions. Obviously, no brick
could be expected to withstand all of these conditions perfectly,
and it would be hardly fair to condemn a brick, for all purposes,
because of its failure to withstand any one of the above condi-
tions. However, it is generally conceded that refractoriness is a
primary requisite of a fire brick for general use and the other
properties, such as ability to resist abrasion or the penetration
of slags, may be said to be of secondary importance. Therefore,
the question of the satisfactory evaluation of refractoriness has
proven the main problem in the testing and classification of clay
fire bricks.

1 By permission of the Director, Bureau of Standards.

AMERICAN CERAMIC SOCIETY. 385

The Measurement of Refractoriness. — In working out a satis-
factory classification of fire bricks based on refractoriness, several
tests for the measurement of this property have been put into
more or less general use. Of these, the three which have been
used the most are, first, the chemical analysis, which gives a
theoretical indication of refractoriness, second, the direct de-
termination of the so-called "softening point," and third, the
determination by an actual "load test" of the ability of a fire brick
to resist deformation under loads at high temperatures. These
tests will be described more in detail later. Another test, as a
measurement of refractoriness, but which has not been used very
extensively in the case of fire bricks, is the determination of the
porosity and volume changes on heating the bricks to different
temperatures. A decrease in porosity and volume indicates the
progress toward vitrification, and when these changes are plotted
in the form of curves, the slope of the curves shows the rate of
vitrification. Over-firing is then indicated by an increase in
porosity and by an increase in volume as the vesicular structure
accompanying over-firing is developed. A classification of fire
clays, based on porosity changes, was first suggested by Purdy,1
who set the limits for the different grades of refractories. These
limits have since been found to be in error but nevertheless the
idea of measuring the refractoriness of a fire brick by its changes
in porosity and volume is evidently a sound one.

Temperature-Porosity Changes as an Indication of Load
Carrying Ability. — The value of the porosity and volume change
determinations is all the more evident when it is realized that the
degree of vitrification, as shown by the changes in porosity, also
represents the amount of softening which must necessarily take
place in the process of vitrification. Marked softening of the
mass of the brick means decreased resistance to deformation. In
other words, the changes in porosity may be said to bear some re-
lation to the resistance to deformation of a brick under load
at high temperatures.

Object of Present Investigation. — From the above con-
siderations it is evident that the determination of the porosity
1 U. S. Ceol. Surv. Bull., 9, p. in.

386 JOURNAL OF THE

and volume changes is very valuable in the study of the funda
mental properties of clay lire bricks and it is only by a better
knowledge <>f the fundamental properties that a completely sati
factory classification of lire bricks can be worked out. The
present investigation was, therefore, undertaken in order to
study the fundamental properties of clay fire bricks by a com
parison of their changes in porosity and volume, upon heating
to different temperatures, with the results of a specific load test
and with the so-called "softening points." Preliminary work
on this investigation was first undertaken under the direction of
G. H. Brown, when connected with the Bureau of vStandards. It
was afterward decided to take up the investigation on a larger
scale, with the cooperation of the Refractories Manufacturers'
Association, which kindly furnished samples of a great number
of brands of fire brick for the tests, and likewise in the interest
of the American Gas Institute.

II. Description of Fire Brick Tests Already Used.

Before taking up the details of the present investigation, it is
well to describe and discuss briefly the most noteworthy tests
which are being used in determining the value of clay fire bricks —
including the tests which have already been mentioned.

Chemical Analysis. — The chemical analysis has been used to
quite an extent in the study of the fundamental properties of
fire bricks and is of considerable value — not only in this regard
but also in explaining the causes of failure in use. The chemical
analysis serves as an indication of the refractoriness by showing
the degree of purity of the clay. Clay of the purest type possi-
ble, i. c, pure kaolin, has a softening temperature or so-called
"melting point," of 17400 C.1 This purest form of clay has a
composition corresponding to that of the mineral kaolinite, being
composed of 46 . 3 per cent silica, 39 . 8 per cent alumina and 139
per cent chemically combined water. In burning, this chemical
water is driven off and the composition becomes 53 . 8 per cent
silica and 46.2 per cent alumina. Clay of this composition is of j
comparatively rare occurrence, owing to the presence, in most

1 Bur. Standards, Tech. Paper, 10.

AMERICAN CERAMIC SOCIETY. 387

ises, of quartz, feldspar, magnesia, lime and iron oxide, in vary-
[g amounts, and which behave as fluxes in lowering the softening
3int of the clay to a greater or less extent. Some attempts have
sen made to apply the chemical analysis directly as a measure
' refractoriness, by means of formulas, but these have not been

any degree satisfactory. The determination of the amount
' these fluxes by a chemical analysis may, however, be con sid-
ed as an aid in determining the refractoriness of the clay, as
ell as serving in some measure to explain the failures of fire clay
rick in use.

A classification of fire bricks based on the chemical analysis has
ten proposed1 — which fixes the limit for the permissible amount

fluxes for No. 1 grade bricks. According to this classification,
le total amount of fluxes must not exceed o. 22 molecular equiva-
nts. This expression for the amount of fluxes is a part of the
npirical formula commonly used in the study of ceramic bodies,
his is calculated from the chemical analysis, by dividing the re-
active percentages by the molecular weights of the oxides,
len dividing each value obtained by the value for the alumina,
he values for the various fluxes are then summed up and the
rmula obtained is of the type ^Si02.iAl203.3/RO. In this for-
ula, determined for fire bricks, the RO equivalent must not
:ceed 0.22 for the No. 1 fire bricks, according to the suggested
assification. In the case of the silicious clay brick (say one con-
king 65 per cent or more of silica) the permissible amount of
ixes is considerably less — since the fluxes are more effective upon
icious material. It is quite evident, then, that the chemical
talysis is of some value in studying the causes of failure of refrac-
ries. However, there are physical tests which require less time

perform than the chemical analysis and yet which indicate
uch more satisfactorily the value of a fire brick in a practical
anner.

Softening Point. — The determination of the so-called "softening
)int" is one of the most widely used tests for fire bricks and is
nerally considered a reliable indication of the refractoriness of
e material. In determining the softening point, small tetra-

1 Bur. Standards Tech. Paper, 7.

388 JOURNAL OF THE

hedra, of the size of the well-known Orton standard higher pyri

metric cones, are molded from a portion of the brick, ground to pal
about an eighty-mesh screen, and mixed with a gum of tragacantl
solution to serve as a temporary bond. These are then set in
suitable plaque of refractory clay material beside a series of stand
ard pyrometric cones and heated in a carbon resistance electa
furnace, or a small pot furnace heated with a gas-blast burne
of the Fletcher type. The "softening point" is indicated by th|
point at which the test cone deforms to the extent that the ti]
is on a level with the base or the cone has swelled excessively
Results are expressed in terms of the standard cone which ha
deformed to the same extent. Temperatures are not usuallj
considered — since it is a well-known fact that the bending of th<
standard cones is affected by the rate of heating. However
very good comparative results are obtained by limiting the timi
of firing in such a test to an hour and a half or two hours.

In classifying fire bricks according to the softening point, com
30 is usually considered the lowest value for No. 1 grade, cone 3<
to 28 for No. 2 grade, and cone 28 to 26 for No. 3 grade
This classification is, no doubt, a proper one but at the same timi
it quite often proves misleading, in so far that the bricks witl
a higher softening point, cone 32 or above, sometimes fail t<
withstand practical conditions of load and temperature whil
silicious bricks, low in fluxes, with a softening point as low a
cone 29, generally stand up well under heavy loads at fairly higl
temperatures. This is explained by the fact that silicious brick
remain quite rigid to within about ioo° C of their softening point
while the brick lower in silica soften through a longer range
Then, too, the softening point determination is misleading i
the case of bricks made from a coarse mixture of very refractor 1
non-plastic clay (flint clay) with an inferior bond clay. Tb
comparatively fine grinding necessary to molding the cone brin^
about a more intimate mixture of the material so that the r<
fractoriness is raised. This objectionable feature of the softei
ing point test is sometimes overcome by using a chip of the origin;
brick — cut on a carborundum wheel to the form and size of tl
standard cones. This, however, is often a difficult task and
practically impossible in the case of a very coarse friable materu

AMERICAN CERAMIC SOCIETY. 389

say nothing of the variation in the results arising from not ob-
ning an average sample from so small a specimen.

rhe Load Test. — What is known as the standardized "load test"
s first used by the Bureau of Standards in an investigation of
y refractories. This test has proven the most practical and
■est means of determining the behavior of fire bricks in use
der high loads and temperatures. The test is generally per-
med upon a brick of standard size and shape, placed on end
a specially designed furnace of about 20" X 20" X 20" internal
nensions. The load is applied to the brick longitudinally,
ough a suitable opening in the top of the kiln, by means of
lighly refractory column on which rests a horizontal beam. On
: furnace used by the Bureau of Standards, this beam rests
on a knife edge on the column and acts as a lever fastened and
justable at one end to keep it level, and the load is applied at
: other end. On a later type of load test furnace, the beam
Dalanced by applying an equal load on both ends and no knife
je is used. The firing, in the case of the Bureau of Standards
nace, is done by means of eight Fletcher gas-blast burners, two
each side, and four in front. In conducting this test, the tem-
ature is raised at a specified rate to the maximum tempera-
e, which is held for one hour and a half. It has been found
it 13500 C is the most suitable maximum temperature in test-
first grade fire bricks. A load of 40 pounds per square inch
; been used, although this is rather severe and one of 25 pounds

square inch is believed a fairer test for No. 1 brick.
Although there are no standard specifications for clay fire bricks
yet, one-half inch contraction in a length of nine-inch brick is
isidered the permissible limit for a No. 1 refractory after
ting under a load of 40 pounds per square inch at 13500 C. All
clay bricks, which have not deformed more than one-half
h in such a test, have proven very satisfactory in use under
rage loads at temperatures not exceeding 13500 C. Hence,
re seems to be no doubt whatever of the practical value of
load test. It may be a little too severe in rejecting bricks
ich might prove satisfactory in use, but this fault is easily
ledied by decreasing the load applied. However, there is

390 JOURNAL <M; THE

one objection to the load test. This is the matter of time re
quired for its performance since only one brick can be tested ii
Mic furnace at a time and each test requires about six hours tim
to perform. This is quite a serious objection when only one tes
kiln is available and the insults on a large number of brick
are needed in a short time.

Cold Crushing Test. This test is a rather valuable one ii
eliminating unsatisfactory bricks from consideration for us
under loads at high temperatures. Obviously, a refractory pos
sesses but a small part of its cold crushing strength when at
temperature of 13500 C, and if a brick has a low crushing strengtl
when cold, it is almost certain to fail in a load test at 13500 C
It has been found that a cold crushing strength of 1000 pound
per square inch, when the brick is tested on end, is about th
lowest permissible value in this regard. Needless to say, a higl
capacity compressive strength testing machine is necessary ii
performing this test.

The Ball Test. — A special form of load test, called the "bal
test," has been put into use by Unger and Nesbitt, of the Carnegi
Steel Company.1 This test was devised in attempting to find 1
substitute for the regular load test and which could be performe(
on a larger number of bricks in a short time. It is claimed tha
the results compare favorably with those of the regular load test
The test is made by heating bricks to 13500 C for one hour i:
a furnace, drawing them out quickly and laying them on thei
side under a long lever in such a manner that a two-inch iron bal
placed on the brick under the lever, is depressed into the fac
as increasing pressure is applied to the lever. A certain load i
applied which is gradually brought up to a maximum in abou
five minutes. The depression made by the ball is measured i
inches. The proper load to give a depression comparable to tt
contraction in a regular load test was determined by experimen
There is no doubt that such a test can be made on a number <
bricks in much less time than is required by the regular load tes
HowTever, there is some objection to it as a regular test for fill
bricks on the ground that cooling takes place sufficiently durii1
1 Proc. Eng. Soc. Western Penn., 32, No. 7.

AMERICAN CERAMIC SOCIETY. 391

1

e test to influence the results. Then, too, the portion of the
ick receiving the pressure is a small part of the whole brick
id the results are apt to be affected by small local defects in
I structure of the brick.

The special tests of fire bricks, such as the determination of the
sistance to slagging action, the resistance to impact and abrasion,
id the test of ability to withstand sudden temperature changes,
,ve little or no connection with the present investigation and
ed not be described here.1 -2

III. Method of Study of Porosity and Volume
langes in Connection with Load Tests and Softening Points.
Determination of Porosity and Volume Changes. — Taking up
e details of the present investigation, the porosity and volume
pages of the fire bricks, for comparison with the results of load
sts and softening points, were determined in the following
Miner: Specimens of each brand of fire brick were cut with a
oad chisel into briquettes, approximately 21 2 X i1 1 X i1 4
ches in size, squared up somewhat on a grinding wheel, and
itably marked with a cobalt paint — made by mixing powdered
bait oxide, one part, and kaolin, three parts, in water. Six
iquettes were cut for each brand. A neater method of cutting
e specimens would have been to use a carborundum cutting
leel. However, the method used was quicker and equally as
iod except for the appearance of the specimens.
The initial porosity was then determined on two of these small
ecimens from each brand — by first weighing the specimens
o. 10 g. after drying at 1 io° C, then determining the wet weight
id the suspended weight after soaking in boiling water under
vacuum, represented by a 24-inch mercury column, for four
iurs. The per cent porosity was then determined from the
ual formula

Wet weight — drv weight

^r— ^7— , , ttt X 100 = per cent porosity.

Wet weight — suspended weight

1 For description of these special tests see Nesbitt and Bell, "Practical
L'thods for Testing Refractory Fire Bricks," Proc. Am. Sue. Testing Ma-
ials, 16, 349, Part 2.

■ Brown, "Methods of Testing Corrosive Action of Slags," Trans. Am.
ram. Sac, 18, 277, ( 1 9 1 6 )

392 JOURNAL OF THE

The volumes of the six briquettes of each brand were then ■
curately determined in a voluminometer of the Seger type, an

soaking the specimens in oil. Duplicate determinations of eac
volume were made to obtain a check within 0.2 cc.

The six briquettes of each brand were then set for re-burnin
in a muffle in a test kiln in such a manner that a specimen f<
each brand could be drawn at regular intervals in the process <
firing. Space was left between each set of draws and be twee
the individual piles to allow the equal heating of all briquette
A set of pyrometric cones and a thermocouple in a porcelain pr<
tecting tube were placed inside the muffle beside the briquette
With this arrangement, the size of the muffle permitted the bun
ing of 84 specimens from 14 brands at one time.

In burning the briquettes, the temperature was brought rapidl
up to 25o°-30o° C and held for three or four hours — in order 1
vaporize the oil in the briquettes (without igniting and leaviu
carbon to influence the results by the reduction of the iron
Over night, the temperature was brought up to 800 ° C, this beit
the highest temperature that could be attained in this kiln wit!
out the use of compressed air — which was not usually availabl
at night. From 800 ° C the temperature was raised rapidly t
12000 C. The temperature was then increased at the rate <
30 ° C per hour from 12000 to 15000 C. The draws were mac
at 500 intervals between 12500 and 15000 C, and wrere cook
slowly by covering with hot sand in a suitable container.

The deformation of the standard cones, as observed durii
the firing of the briquettes, was as follows:

Cone 13 down at 1350° C

Cone 15 down at 1450° C

Cone 17 down at I 0 „

Cone 18 half down at J

As usual, a cone was considered down when the tip w
on a level with the base. It may be well to note here that com;
when used in a burn of this kind, are retarded slightly, owing
the cooling unavoidably given them in making the draws.

After the burning of the briquettes, the porosity of each w
determined by the same method used in determining the init
porosities. It may be said in regard to the method of soaki

AMERICAN CERAMIC SOCIETY. 393

: specimens, that placing them in boiling water under vacuum
four hours gave more consistent results than the usual method
simply boiling in water. It was found that results could be
^eked exactly on the same specimen by the vacuum method,
ereas, results by the boiling test alone were found to be some-
les as much as three points lower than by the vacuum method,
["he final volumes of the briquettes were determined in the
uminometer in the same manner as the initial volumes —
Dlicate determinations on the same briquettes being checked
0.2 cc. The volume change was computed in terms of the
:ial volume. The volume changes and porosities at the differ-
temperatures were plotted in the form of curves for each
nd of brick.

Method of Making Load Test. — The regular load test was made
a full-size specimen of each brand (standard size bricks used)
ploying a load of 40 pounds per square inch at a tempera-
e of 13500 C for an hour and a half. The load of 40 pounds

square inch was considered best for testing first grade fire
:ks at the time the tests were made, although one of 25 pounds

square inch is now generally considered as best.
Tie kiln used for the load test was the one which has already
n described. Special care was used during these tests to keep

beam level at all times in order to prevent eccentric loading
the specimen. The usual rate of firing up to 13500 C was
owed. It is given in Table 1.

Table i . — Schedule of Firing in Load Test.

Time.

Temperature.
Degrees C.

Time.

Temperature
Deerees C.

ours. Minutes.

Hours.

M

inutes.

35

200

2

45

1 1 60

3D

37°

3

1200

45

520

3

15

1240

1

670

3

3fJ

1270

I 15

800

3

45

1300

1 30

880

4

1320

1 4.5

960

4

15

1340

2

I020
IO7O

4

30

1350

2 *5

2 30

I I20

1350°

C held for

one

hour and a

half.

394 JOURNAL OF THE

riic thermocouple, in ;i porcelain protecting tube, used i:
measuring the temperature, was placed within one half inch I
the brick. Cones were also placed close to the brick as a died
on the firing although the couples and instruments used in tin
test and for the porosity determinations wen calibrated again
standards, from time to time, to insure accuracy in measuring tin
temperatures.

When the investigation was first started, the plan was to tnai
duplicate determinations by the load test and to take the aver
age results for comparison with the results of the porosity an<
volume change determinations. However, the variation in initia
porosity on different specimens of the same brand of brick i
sometimes quite considerable, due, in most cases, to a differeno
in burning in manufacture. It was therefore decided to deter
mine the initial porosity on several of the full-size bricks of eacl
brand, in the same manner as was done with the briquettes, anc
to use the brick for the load test which was found to have tin
nearest porosity corresponding to that of the briquettes used it
the other test. By so doing, it seemed that a good comparisot
of the results of the load test with those of the porosity and vol
ume change determinations could be made without making dupli
cate determinations by the load test on each brand. Under th
circumstances, it did not seem worth while to take the time fo
more than one load test on each brand — considering the larg
number of brands tested.

Method of Determining Softening Points. — The softening
point determinations were made by grinding up portions of th
brick to pass a 6o-mesh screen and molding with gum of tragi
canth into cones as nearly as possible the exact size of the Orto
higher pyrometric cones. These were set in a plaque of suitab!
refractory clay material with a series of pyrometric cones — -i
the usual manner of determining softening points — and fired i
a small pot furnace, using a Fletcher blast burner with natur
gas and compressed air. The time of firing in each test w;
limited to an hour and a half to two hours. The softening poii
was determined by the point at which the cones bent in the pr
scribed manner and the result was expressed in terms of the corr

AMERICAN CERAMIC SOCIETY. 395

ponding standard cone. Duplicate determinations of the soften
ig point were made on two bricks of each brand. Practically
,11 of the results agreed within one-fourth of a cone and in no case
lid duplicate determinations on two bricks differ more than one-
lalf cone.

IV. Discussion of Results.

A study of the data obtained shows some interesting relation-
hips between the porosity and volume changes and the results
)f the load tests. In nearly all cases of brick which successfully
withstood the load test — showing a deformation of not more
han 5 . 55 per cent — there is comparatively little volume change
»y either expansion or contraction up to i425°C, while a consid-
rable number of the brick which failed in the load test show
■cry appreciable volume changes. Similarly, the porosity de-
rease in the case of brick which pass the load test is not large
yhile many of those which failed show considerable decrease in
•orosity at some point below 1425 ° C. In many cases, over-
mrning of the brick of poorer grade is distinctly evident from
he volume and porosity changes. The sudden and more or less
>ronounced expausion at the lower temperatures of firing, in the
ase of some of the bricks, is certain evidence of over-burning which
vas confirmed by the results in the load tests. An abrupt in-
■rease in porosity at the same or at a slightly higher tempera-
ure is usually noted in such cases. Invariably, such brick
ailed in the load test. Bricks which show either marked volume
:hange or a considerable decrease in porosity also failed to with-
stand the load test.

It would seem, then, that the volume and porosity changes
)f clay fire brick might serve in some measure as a criterion of
heir ability to withstand load at high temperatures. A care-
til study of the following data for various burning tempera-
ures shows some interesting points in this connection :

1350 ° C: Of 26 brick which passed in the load test, 23 show
lot more than 1 per cent volume change, 3 show more than 1
:>er cent. Of 35 brick failing in the load test (showing more then
5.55 per cent deformation), 16 show not more than 1 per cent
/olume change and 19 more than 1 per cent.

{96 JOURNAL OF THE

( )l" 26 brick passing the load test, 25 show no1 more than 3 per

cent porosity decrease, 1 more Uian 3 per cent. Of 35 failing in

the load tests, [3 show not over 3 per cent and 22 more than 3
per cent porosity decrease.

Combining these criteria, 33 of 61 brick show not more than
1 per cent volume change nor more than 3 per cent porosity
decrease — 23 of these passing and 10 failing the load test. Of
those showing more than 1 per cent volume change or more than
3 per cent decrease in porosity, 3 passed and 25 failed the load

tests.

1400 ° C: Not over 2 per cent volume change was shown by
23 brick which passed and 15 which failed in the load test. Of
23 brick which showed over 2 per cent volume change, 3 passed
and 20 failed in the load tests.

Of 36 brick showing not over 5 per cent decrease in porosity,
22 passed and 14 failed in the load test. Of 25 showing more than
5 per cent porosity decrease, 3 passed and 22 failed in the load
test.

< )f the brick which showed neither a volume change greater
than 2 per cent nor a porosity decrease exceeding 5 per cent,
22 passed and 10 failed in the load test.

On the basis of a volume change not exceeding 3 per cent at
14000 C (amounting to about 1 per cent linear expansion or con-
traction), 24 passed and 19 failed in the load test. Of those
exceeding 3 per cent volume change, 2 passed and 16 failed the
load tests.

Excluding brick which showed more than 3 per cent volume
change or 5 per cent porosity decrease, 22 passed and 10 failed
in the load test. Of those which showed either more than. 3 per
cent volume change or 5 per cent porosity decrease, 4 passed
and 25 failed in the load test.

1425 ° C: Of 38 brick showing not more than 2 per cent vol-
ume change, 23 passed and 15 failed the load test. Of those
showing more than 2 per cent volume change, 3 passed and 20
failed the load tests.

Of 34 brick showing not over 5 per cent porosity change, 21
passed and 13 failed the load test. Of those showing more than

AMERICAN CERAMIC SOCIETY. 397

5 per cent porosity decrease, 6 passed and 21 failed in the load
tests.

Of the brick which neither changed more than 2 per cent in
volume nor 5 per cent in porosity, 20 passed and 7 failed in the
load test. Of those changing more than 2 per cent in volume
or 5 per cent in porosity, 6 passed and 28 failed in the load test.

It is apparent that the porosity and volume changes of clay
fire brick, when burned at some temperature between 13500 and
1425 ° C, offers, in some degree, a criterion of their ability to with-
stand the load test. The bricks passed or rejected by the sug-
gested limitations in porosity or volume changes at the various
temperatures are nearly the same in all cases. This consistency
indicates that the apparent relation between these changes and
the ability of bricks to withstand the load test is more than a
chance one and that certain limitations in this respect might well
be used as a means of rejection of a large number of bricks which
would undoubtedly fail in that test. The use of either porosity
or volume changes alone offers a fairly satisfactory criterion.
Taken together, the result is even better — since a number of
brick failing in the load test can pass one or the other specifica-
tion as regards volume or porosity changes but few of the better
bricks fail to pass both.

For practical purposes, a temperature should be selected which
is not too high to be readily attained under ordinary conditions,
but high enough that the limitations in porosity or volume changes
need not be too close to be easily measured. The use of linear
measurements of expansion or contraction rather than volume
measurements, would be extremely desirable from the standpoint
of simplicity. From the results here obtained, a temperature of
14000 C would seem most satisfactory since the permissible limit
in volume change may be made as high as 3 per cent — this prac-
tically amounting to a linear contraction or expansion of 1 per
cent. A limit of 5 per cent in porosity decrease is also quite
satisfactory. At this temperature also the effects of over-burn-
ing in the inferior materials is likely to become sufficiently evi-
dent to be readily detected. Nearly all of the bricks which suc-
cessfully withstand the load test could meet these specifica-
tions while a majority of those that fail under load would be elim-

398 |« >URNAL OF THE

inated by excessive volume or porosity change, or both. It maj
be noted thai in all cases brick which showed a porosity decrease
exceeding ,s per cent, or a volume change exceeding 3 per c< at,

showed a deformation in excess of s per cent in the loud test
close to the permissible limit.

In the case of clay tire brick, an increase in volume is especially
significant, as it represents evidence showing that over-firing has
taken place -due to a comparatively high content of fluxes. It is
important, therefore, that the permissible expansion be rigidly
limited to a definite value which must be lower than that for
the allowable contraction.

In the case of some of the brick failing in the load test — -due,
to failure of the bond there were no marked changes in either
volume or porosity. Generally, such brick have a very low, cold
crushing strength — as in the case of one brick which failed un-
der a load of 364 lbs. per square inch and another which crushed
at 385 lbs. per square inch. The minimum allowable cold crush-
ing strength for a No. 1 refractory is generally placed at rood
lbs. per sq. inch.

There is apparently little relation between the softening points
and the volume and porosity changes or with the results of the
load test. Some of the brick, showing high softening tempera-
tures, failed to withstand this test while several which softened
below cone 31 showed less than 5 per cent deformation. The
latter were probably silicious in character, containing more than
65 per cent Si02. In some cases where brick failed, the high
softening point is probably due to the intimate mixture of the
inferior bonding clay with the more refractory flint clay in the
preparation of the test cones — the material being ground to 60-
mesh size. In all cases, brick showing softening points lower than
cone 28 failed completely in the load test. It may be con-
sidered then that cone 28 should be the minimum softening tem-
perature of a No. 1 refractory of either silicious or non-silicious
character.

In Fig. 1 are shown porosity and volume change curves for six
bricks of different characteristics which successfully withstood
the load test. These have been selected as fairly typical of the

AMERICAN CERAMIC SOCIETY.

399

ricks of this class. Contraction is indicated by negative volume
liange, expansion by positive values.

/Y /Jo* /JSo '4>» rf& /■*><

\

*

&

*

^J

M

3

\

■X

t

\

X

' /&j /££> /4>o /4ji> /.

Fig.

In the curves for brick No. i a gradual expansion, accompanied
iy a gradual decrease in porosity, is noted. The low initial
torosity is characteristic of the stiff -mud brick. The gradual ex-
>ansion would seem to indicate a silicious character, but this

400 JOURNAL OF THE

seems unlikely in view of the high softening point, cone ,^2. The
result of the load test was a deformation of 4.47 per cent.

Brick No. [3 shows gradual contraction beyond 1350° C-
aecompanied by an increased rate of porosity decrease above
i425°C. This is also a stiff-mud brick with low initial porosity
and a softening point of cone 318 1. The deformation in tin
load test was slight, only 1 .06 per cent.

Brick No. 18, made by dry pressing, is apparently silicious in
character— as indicated by the slight expansion on heating and
the very slight porosity changes. The slight deformation in
the load test, o . 1 1 per cent, and the softening temperature of
cone 29V4, which is too low for a non-silieious material of this
quality, also indicate a high silica content. The chemical analysis
shows 80.56 per cent SiC>2 and a very low content of fluxes.

Brick No. 20, hand made from a mixture of flint and plastic
clays, represents a high clay brick of excellent refractoriness.
The porosity curve shows the characteristics of flint clays — in
the slight changes up to 1425 ° C. No appreciable volume change
occurs below 14500 C. The softening point was cone 32 and the
deformation in the load test was only 1 .62 per cent.

Brick No. 45 showed a deformation of 5 .33 per cent in the load
test — close to the allowable limit. The large amounts of contrac-
tion and porosity decrease would render the quality of this brick
doubtful in the light of the suggested specifications in this re-
spect. The regularity of the porosity decrease and contraction
indicates a gradual and continuous fluxing action but with no
indication of over-burning. The softening point of this material
was cone 3 1 .

Brick No. 48 shows a rapid porosity decrease above 13000 C.
The volume change is slight up to 1425 ° C, where over-burning
is indicated by the subsequent expansion. While the deforma-
tion in the load test (4.47 per cent) fell within the limit, the
porosity change exceeds 5 per cent. The ability of this brick to
withstand a long-continued heat treatment at high temperatures
would be somewhat doubtful — due to its over-burning tendencies.
The softening point was at cone 31 ! 4.

In Fig. 2 are shown curves for some of the brick which failed
to pass the load tests. Brick No. 26, with a deformation of 12.6

AMERICAN CERAMIC SOCIETY.

401

cent, shows a gradual expansion above 13000 C, becoming
id above 14000 C. The increase in porosity at 1425° C in-
ites that this is probably due to the gradual development
I vesicular structure by over-burning.

'J<V life 7Wt> /4fi> 73a*

dur/j/ng 7?/»fie/-afure //7 °6.

/y /Jee /AT* /for "P& /Jar

$vrV7/hfr 7e/rr/>f/-a/Wre //* V

Fig. 2.

Jrick No. 38 shows marked but gradual fluxing in its rapid
rease in both porosity and volume. It is evidently not of
:ious character and of only fair refractoriness. The softening

,,,.■ JOURNAL OF THE

point at cone 301 1 and the deformation in the load test (I
per cent) confirms this.

Brick No. 41 begins to over-burn at 13000 C, as indicated
the expansion and by the break in the porosity curve at a soi
what higher temperature. The high softening point, cone 3I
is apparently the result of the fine grinding of the material
preparing the test cones. The failure to pass the load test
probably due to an inferior bond clay in the mixture.

Brick No. 49, also made from a mixture of clays, shows CM
burning by the expansion above 13000 C and the break in
porosity curve. The softening point is at cone 31V4 and the
formation in the load test was 9.37 per cent.

Brick No. 8 is typical of a number of those which failed c(
pletely in the load test. It has a low softening point (cone 27
and shows marked over- burning above 13000 C, as indicated
the rapid expansion and by the increase in porosity. The
cided contraction and decrease in porosity when burned at 131
C indicate a low temperature of burning in manufacture.

V. Suggestions for the Practical Application of Porosity a
Volume Changes of Clay Fire Brick.

The determination of the porosity or volume changes of clay
brick, burned at some temperature between 1350 ° C and 1425 c
affords a means of classifying a certain proportion of brick wl
would fail in the load test. The limit of 3 per cent in volt
change (or 1 per cent in linear contraction or expansion) an
per cent decrease in porosity at 14000 C, does not eliminate
brick which cannot withstand such a test, but materially redi
the number which would require such testing. In this resp
it would be of considerable advantage in comparing the qua
of a large number of samples — since the porosity and voh
changes may be determined readily from a single burn in a
kiln. To make standard load tests on a number of brick
quires considerable time and expense even if several furn;
are available.

The determination of volume changes is comparatively sin
when porosities are also determined — since the difference
the wet and suspended weights of the specimens affords a meas

AMERICAN CERAMIC SOCIETY. 403

t of volume practically as accurate as determinations by
ns of the voluminometer.

lie porosity and volume changes in burning may also be used
mtageously as a means of determining the uniformity of
ity of a lot of bricks of a single brand. The behavior of such
cs should be fairly uniform if they are of equal quality. Such
ications of these measurements would be of value to the manu-
lrer as well as to the consumer.

study of the curves for porosity and volume changes for a
s of burning temperatures offers a means of determining, in
! measure, the causes of failure of certain brick in the load

If this is due to under-burning, resulting in excessive de-
ation under load — due to shrinkage — a considerable decrease
^rosity and volume will be found at the lower burning tem-
tures. Clays of inferior quality, which over-burn, will be
:ted by a marked expansion at some point in the burning,
lly accompanied at a somewhat higher temperature by an
ase in porosity — due to the development of a vesicular struc-

Clays high in fluxes, but which do not readily develop a
ular structure, show gradual but marked contraction and
;ase in porosity as the burning temperature is increased.

VI. Summary.

Brick which are capable of withstanding a pressure of 40
per square inch at 13500 C generally show slight changes in
me or in porosity when burned at temperatures up to 1425 ° C.

The greater number of the brick which failed to pass
oad tests show rather marked changes in volume or in porosity
•me temperature below 1425 ° C.

Brick which show distinct over-burning by pronounced
nsion at temperatures below 14000 C invariably fail in the
test. The adoption of a definite limit for the permissible
nsion within the given temperature range is particularly
•rtant for detecting inferior clay refractories. The limit for
nsion should be lower than that for allowable contraction.

The changes in volume and in porosity of brick burned at
: temperature between 13500 C and 14250 C serves, in a
sure, as a criterion of their ability to pass the load test.

,,,, WlHkk'.W CKKAMIC SOCIKTY.

5. Most of the brick which show a porosity decrease not e:
ceeding 5 per cent and a volume change not exceeding 3 per ofl
(approximately 1 per cent in linear dimensions), when burnt
at 14000 C will pass the load test.

6. Brick which show a decrease in porosity exceeding 5 p<
cent or an expansion or contraction in excess of 3 per cent 1
volume (1 per cent in length) at 14000 C failed to pass the loa
test in nearly all cases.

7. The use of limiting porosity and volume changes for cfl
fire brick burned at 14000 C would serve as a means of eliminate
from consideration a large number of brick which fail in the loa
tests.

8. Brick which fail in the load test — due to failure in the bond-
may not show marked changes in volume or porosity in burnir
but often show very low cold crushing strength.

9. No definite relationship seems to exist between the softei
ing point of a fire brick and its ability to withstand load at hig
temperatures. However, all brick which softened below cor
28, whether silicious in character or not, failed completely in tl
load test. It seems advisable, then, to specify cone 28 as tl
minimum softening point for any clay fire brick. It is probab
that brick containing less than 65 per cent Si02 should have
minimum softening point of cone 31.

In conclusion, the writer desires to express his appreciatic
of the assistance rendered by Mr. D. W. Ross in conducting tl
load tests of this investigation and to thank Mr. A. V. Bleii
inger for valuable suggestions in carrying out the work.

HE CALCULATION OF THE "RATIONAL ANALYSIS" OF

CLAYS.

By Henry S. Washington.

Introduction.

There are comparatively few clays used in the ceramic in-
istries which consist of pure kaolin, Al2O3.2SiO2.2H2O, or of
ose other closely similar hydrated aluminum silicates, which,
llectively, have been called "clay substance" and by other
nilar names.1 By far the greater number of clays contain other
inerals as well; of these, the most important are quartz, feld-
ars, and micas. The two most common micas, the purely
•tassic muscovite and the ferro-magnesian biotite, are, how-
er, not present in large amount in any but a few clays that are
commercial value, particularly for the finer ceramic purposes.
;sides the above-named minerals, numerous other minerals have
en observed in clays,2 but these, almost without exception,
e present in such small amounts in clays of any notable com-
ercial value that they may be regarded as practically negligible.
Only four minerals, then, kaolin, quartz, feldspar, and mica,
ed be considered for the present. As each of these influences
e "burning" and other properties of a clay and of the finished
oducts, it is of importance to be able to estimate the relative
nounts of each present in any given clay, so that the proper
ixtures may be made or the proper temperature and other
nditions may be controlled. The ceramist endeavors to accom-
ish this by what is known technically as the "rational" analysis.3

1 K. Langenbeck, "The Chemistry of Pottery," Easton, 1895, P- ll'<
W. Mellor, "A Treatise on Quantitative Inorganic Analysis," London,
*3i P- 656; H. Ries, "Clays, Their Occurrence, Properties and Uses," New
>rk, 1914, p. 5.

- Mellor, "Quant. Anal.," p. 656; Ries, "Clays," pp. 56ff.

1 As will be seen, I am quite in agreement with Mellor ("Clay and Pottery
dustries," 1914, I, p. 109, in his low opinion of the "rational" analysis
d in his view that the adjective is misapplied.

p6 JOURNAL l »l; THE

The "rational" analysis aims at the estimation of the amounts
of the different minerals that arc present in a clay; the object
of the "ultimate" analysis is to determine the amounts of the
chemical constituents of the clay, quite irrespective of the mineral
forms in which they may occur.

The "rational" analysis seeks to attain its object by treatment
of the clay with various strong reagents, particularly hydro-
chloric and sulphuric acids and alkali hydroxides. The methods
and details of manipulation differ somewhat,1 and the results of
the ultimate analyses are frequently used in conjunction with the
data so obtained.

Following Mellor, the essential features of the "rational" analysis
are: digestion of the clay with hot, concentrated sulphuric acid
and subsequent washing alternately with alkali and hydrochloric
acid, to remove the "clay substance;" this may be preceded by
treatment with hot alkali to remove soluble silica, or with dilute
hydrochloric acid, to remove the carbonates; and it may be fol-
lowed by evaporation of the residue with hydrofluoric acid to
remove quartz and other silica, and determination of the alumina,
which is used as a basis for the calculation of the feldspar present.

There is great divergence of opinion, even among ceramists
themselves,2 as to the value of the "rational" analysis, a crude
form of which was in use as early as the end of the eighteenth
century.3 It is unnecessary, within the scope of this paper, to
discuss the details of the various modifications which "rational"
analvsis has assumed, and it must suffice to point out a few general
truths that are applicable to the methods in all forms of "rational"
analysis, and that are self-evident to the mineralogist or to the
chemist who has had experience in the quantitative chemical
analysis of silicates.

1 For descriptions of the methods of "rational" analysis of clays see H. A. j
Seger, "Collected Works," Easton, Penna., 1902, I, p. 50; Binns, Zimmer
and Orton, Trans. Am. Ceram. Soc, 2, 221 (1900); H. Bollenbach, Sprechsaal,
1908, No. 25, p. 340; No. 26, p. 351; Mellor, "Quant. Anal.," 1913, p. 658;
Ries, "Clays," 1914, p. 75; A. S. Watts, Bur. Mines Bull., S3, 38 (1913)-

2 For some discussions of the value of "rational" analyses, see C. F.
Binns, Trans. Am. Ceram. Soc, 8, 198 (1906); R. C. Purdy, Ibid., 14, 359
(1912); J. W. Mellor, "Clay and Pottery Industries," 1914, P- 109.

3 Mellor, "Quant. Anal.," p. 657.

AMERICAN CERAMIC SOCIETY. 407

Some, though comparatively few, researches have been made
lto the action of the various reagents used in "rational" analysis
n the different minerals present in clays, particularly kaolin,1
uartz, feldspar, and mica. These researches have neglected to
ike into account all of the factors that may have a great con-
-olling influence on the ultimate effect of the reagents.

Foremost among these factors is the chemical composition of
ic mineral treated. The "feldspar" in a clay may be the potassic
rthoclase, which is, other conditions being alike, only slightly
ttacked by strong acids. It may be one of the series of the
lagioclases, or soda-lime feldspars, which in solubility relations
ary continuously from the purely sodic albite, which is about as
isoluble as orthoclase, through the sodi-calcic labradorite, which
somewhat soluble, to the purely calcic anorthite, which is
iadily soluble, even in rather weak acid. The different micas,
[so, vary in solubility, muscovite being attacked with some
ifficulty, while biotite is much more readily dissolved, the
ifferent varieties of this also varying in their behavior. Kaolin
self varies in its resistance to reagents, both acid and alkali,2
ith its origin, the kind and concentration of acid, and other
Hiditions. The different varieties of silica which may be present
1 clays vary in their resistance to different reagents.

Another important factor is the size of grain, which influences
ery much the extent and the rapidity of attack or solution of

mineral by reagents or solvents. It is well known that the
nailer the particles of a substance the more rapidly it dissolves,
le relation depending on the truth that the area of a sphere
iminishes as the square, and the volume or mass as the cube, of
le radius. The difference in rate of solution with size of grain

accentuated, or the solution-rate gradient becomes steeper,
articularly with ordinarily "insoluble" minerals, when the
imensions of the particles are of the orders of magnitude of those

1 This term is used here to designate any of the clay-like, hydrated,
uminum silicates, whether crystallized like kaolinite, or "amorphous" like
ay substance. It is the general group name used by mineralogists.

2 See H. Stremme, in Doelter, "Handbuch der Mineralchemie," 1914,
. P- 77-

|os JOURNAL OF THE

that make clays.' Mellor' has made some- experiments along this
line on feldspars and micas, with instructive results, as he adopted

for his work the treatment used in "rational" analysis. On
treatment of a sanidine (orthoclase) with sulphuric acid, for
example, he found that only 0.34 per cent was dissolved if the
average diameter of grain was 0.165 mm., while 15.64 per cent
was dissolved if the average diameter was 0.032 mm.:i

< >ther factors that will seriously influence the results are: the
concentration of the reagent, the length of time during which the
material is treated, the temperature at which the mixture is
treated, and whether this is attained suddenly or gradually or is
uniform or varied during the treatment. All of these condi-
tions, and some others, will affect, and generally affect very
seriously, the results that are obtained by "rational" analysis.

The various factors enumerated have such an important bear-
ing on the operation of all the modifications of "rational" analysis
that its methods are scarcely susceptible of yielding satisfactory
duplicate results, except by chance, and even satisfactory duplica-
tion of results cannot be regarded as evidence, still less as proof,
of their correctness, when the methods are seriously faulty. One
can only consider, then, that the methods of "rational" analysis
can furnish little but approximate and uncertain, and probably
more or less erroneous and misleading, information. It would
appear, therefore, that such an analysis is anything but rational.4
It would be better to call such an analysis "modal," as it aims at
determining the "mode" or mineral composition.

It must furthermore, however regretfully, be pointed out that
many writers on, and workers in, ceramics have failed to utilize
even the unsatisfactory data of the "rational" analysis with proper
understanding of the chemical and mineralogical principles that
are involved. Examples of this are numerous in the literature,

1 It is largely because of this that one must be cautious in applying to
clays the data on solubility given in Dana and other text-books; these are
rough and are based on particles of much larger size.

2 Mellor, "Quant. Anal.," p. 662.

3 See also Langenbeck, "The Chemistry of Pottery," p. 9.

4 Cf. Mellor, "Quant. Anal.," p. 674; "Clay and Pottery Industries," p.
1 10.

AMERICAN CERAMIC SOCIETY. 409

and some will be given later, but it may be said here that the ap-
plication of scientific methods to practical ceramics demands a
greater appreciation and understanding of scientific methods than
seems now to obtain.

In contrast with the uncertainties of the "rational" analysis,
the "ultimate," or what is better called the "chemical," analysis
can be carried out with ease and great accuracy by the usual
methods for the analysis of silicate rocks.1 For a clay the usual
chemical analysis takes little more time than the rational. Its
results, stated in terms of the constituent oxides, may be readily
reduced to mineralogical terms by a simple process of calculation
that is well known to students of igneous rocks and that will
yield information as to the mineral composition of a clay that is
far more accurate and certain than the information furnished
by the usual "rational" analysis, even when the latter is properly
executed and the calculations are correctly made.

It is the chief purpose of this paper to suggest to ceramists
this simple and expeditious method of calculating the mineral
composition of a mixture of silicates from the chemical analysis,
and to describe a modification adapted to the mineralogical
characters of clays. The method suggested and to be described
is only a special application of the principles and methods of
calculation of the so-called "norm" of igneous rocks.2 The
"norm" is the expression of the chemical composition in terms of
certain mineral molecules which are assumed to be, and which,
in the great majority of cases, are, actually present in the rock.
The actual mineralogical composition of a rock is called the
"mode," while the rock is classified by the standard molecular

1 W. F. Hillebrand, "The Analysis of Silicate and Carbonate Rocks,"
U. S. Geol. Surv. Bull., 422 (1916); J. W. Mellor, "Quantitative Inorganic
Analysis," 1913; H. S. Washington, "Manual of the Chemical Analysis of
Rocks," 1918.

2 Cross, Iddings, Pirsson and' Washington, /. Geol., 10, 604, 642;
also "Quantitative Classification of Igneous Rocks," Chicago, 1903, pp.
147, 1 36; G. I. Finlay, "Introduction to the Study of Igneous Rocks,"
New York, 1913, pp. 150, 154; J. P. Iddings, "Igneous Rocks," New York,
1909, I, pp. 419, 433; H. S. Washington, "Chemical Anyalysis of Igneous
Rocks," U. S. Geol. Surv., Prof. Paper, 99, 1151-1182 (1917); also separate
publication of this portion of Prof. Paper, 99, p. 1918.

Ho JOURNAL OF THE

composition expressed in the "norm." With clays, in general,
the problem and the calculation art' much simpler than they are

with rocks since in most clays we have to consider and deal
with only three essential minerals, namely, kaolin, quartz, and the
feldspars; the micas are of secondary importance and will be taken
into account later.

Some of the ideas of the method and some of the data of the
calculation have been suggested and used to a limited extent by
various writers,1 and in an interesting paper Staley2 has sug-
gested the application of the principles of the norm to the calcula-
tion of ceramic mixtures, especially to glazes and enamel mixtures.

vStated briefly, the principle consists in utilizing our knowl-
edge of the various molecules which the constituent oxides are
capable of forming, and which they actually do form, and in
considering the minerals present in the rock or clay as made up
of such molecules, the actual minerals into which the molecules
combine being dependent on the conditions and being in accord-
ance with our knowledge of the occurrence of minerals in rocks.

In igneous rocks we are dealing with solidified solutions, during
whose solidification and consequent crystallization as a mixture
of definite minerals, certain laws of so-called "affinity" of the
various basic oxides for silica and alumina come into play and de-
termine the actual mineral composition — which may vary some-
what with the varying conditions that obtain during the solidifica-
tion.

In clays and similar bodies of secondary origin, on the other
hand, we are dealing with the remnants of the alteration, de-
composition, and disintegration of these minerals. In these
remnants we see not only the effects of the laws just mentioned,
but also the action of others which seem to pertain especially
to those phases of metamorphism that are known as "alteration"
and "weathering," among the products of which are the clays of the
potter.

1 E. g., E. R. Buckley, "Report on the Clays ofWisconsin," p. iooi; C. F.
Binns, Trans. Am. Ceram. Soc, 8, 203 (1906); J. W. Mellor, "Quant. Anal.,"
pp. 660, 671, 674.

2 H. F. Staley, Trans. Am. Ceram. Soc, 13, 126 (1911); Prof. Staley's
suggestion only came to my attention when the present paper had been al-
most completed.

AMERICAN CERAMIC SOCIETY. 411

Let us now consider the example which especially interests us
here, and to which we shall confine our attention henceforth: a
clay, whether a kaolin, or a porcelain-, ball-, fire-, brick-, stone-
ware-clay, or other variety.

The chemical ("ultimate") composition of this is expressed
as made up of silica (SiOL>), alumina (AI2O3), ferric oxide1 (Fe-..* >;),
magnesia (MgO), lime (CaO), potash (K20), soda (Na20), water
(IIjO), with sometimes small but possibly notable amounts of
titanium dioxide (Ti()2) and carbon dioxide (CO2), and very
small and negligible amounts of other constituents.

In mineralogical or "modal"2 composition, all clays are essentially
mixtures of kaolin = Kl, Al2().j.2Si02.2H20,:i quartz = Q, and the
feldspars, orthoclase = Or (K2O.Al203.6Si02), albite = Ab
(Na2O.Al2O3.6SiO?), and anorthite = An (Ca0.Al203.2Si02) ; and,
in some clays, small amounts of muscovite = Mu (K20.2H;>0.-
3Al203.6Si02). For the moment, the presence of muscovite and
of other minerals will be disregarded, so that the principles may
be more readily explained.

The problem before us then is to express the chemical composi-
tion of a clay in terms of the minerals kaolinite, quartz and the
three feldspars. Their formulas represent the molecular com-
position; thus, kaolinite is made up of one molecule of alumina
(AI2O3), two of silica (Si02), and two of water (H20). From the
analysis of a mineral the formula is found by reducing the con-
stituents to molecules and finding the ratios of these. We analyse
a pure kaolinite and, neglecting the unavoidable slight impurities,

1 Ferrous oxide (FeO) is not considered here, as it is seldom determined
in clay analyses and may be regarded as oxidized to ferric oxide during the
burning.

2 This term, used to express the actual quantitative mineral composition
of a rock, is borrowed from the nomenclature of the Quantitative System of
classification of igneous rocks. It corresponds to the "rational" composition
of the ceramist, but it must be remarked that the chemical composition is
just as rational as the mineralogical.

3 This, the well-established formula of kaolinite, may be taken as that
of the clay substance, which is represented by such minerals as kaolinite,
halloysite, newtonite, cimolite, and other similar substances of more or less
doubtful and possibly heterogeneous composition. They represent the
aluminous molecule corresponding to the magnesian serpentine, and differ
chemically only in the amount of water content.

4i2 Ji lURNAL OF THE

we divide the percentage amount of alumina by its molecular
weight (102), that of silica by its molecular weight (60) and thai

of water by its molecular weight (18). It is found that the
quotients, which represent the relative amounts of the mole
cules, are in the ratio 1:2:2, and consequently we write the
formula of kaolinite as Al2Os.2SiO2.2H2O. The molecular weight
of kaolinite is thus 102 -f- 120 + 36 = 258; that of quartz is 60,
that of orthoclase is 556, that of albitc is 524, and that of anorthite
is 278.

We have then an analysis of a clay composed "modally," as
stated above, and to find its mineral composition from these
chemical data we reduce the percentage amounts of the various
constituents to molecules by dividing each by its molecular
weight; in this case Si02 by 60, AI0O3 by 102, H2O by 18, K20 by
94, Na20 by 62, and CaO by 56.

Studying the formulas of the minerals under consideration,
we see that in each of the feldspars one molecule of K20, Xa20,
or CaO is present; A1203 is present in the three feldspars, and in
each in equal ratio (1:1) with the K20, Na20, or CaO, and is
present as one molecule as well in kaolinite; Si02 is present in all,
but in different ratios to the other oxides, making up all of the
quartz, being sixfold as much as K20, Na20, and A1203 in ortho-
clase and albite, and twice as much as CaO or A1203 in anorthite
and kaolinite; H20 is only present in kaolinite, where it is equal
to the Si02 and twice as much as the A1203, and it may also be
present, uncombined chemically, as moisture.

The matter now resolves itself into algebra of a very simple
sort. For convenience and simplicity, at the start we shall assume
that orthoclase (K2O.Al203.6Si02) is the only feldspar present.
To the (molecular) K20 present we assign an equal (molecular)
amount of AI0O3,1 and six times the (molecular) amount of Si02,
to form the orthoclase molecule. As free alumina does not exist
in clays, the rest of the A1203 must go into kaolinite; so of the
Si02 we set aside for this mineral molecule an amount out of what

1 If we take albite and anorthite into account, we should assign, of the
Al20.i, amounts equal to those of Na-jO and CaO to form these minerals,
respectively, and of the Si02 six times that of the NaaO to form albite and twice
that of the CaO to form anorthite.

AMERICAN CERAMIC SOCIETY. 413

s left over from the feldspar equal to twice the amount of the
esidual A120.;; we set aside also an equivalent amount of H20.
Phis uses up all of the K20 and A1203 present, and there is left
ixcess Si02 and probably H20. The Si02 is the quartz molecule,
tnd the excess H20 represents moisture.

We now have the molecular amounts of the orthoclase, kaolinite,
tnd quartz, and to reduce these to mineral percentages we simply
lave to reverse the operation of obtaining the molecules; that is,
ye multiply the molecular amounts by the molecular weights.
Phis can be done by multiplying the molecular amount of each
issigned portion of each individual oxide by its proper molecular
weight; but it is easier and more expeditious to multiply the
noleeular weight of each mineral by the figure that expresses the
nolecular amount of this mineral. This figure is simply the
noleeular number of that constituent oxide of which only one
nolecule is present in the mineral. Thus, for kaolinite it would
•e the molecular amount of AI2O3 present in kaolinite, for ortho-
lase the molecular amount of either K20 or Al20,3 present in
rthoclase, and for quartz the molecular amount of the residual
ii02. The results of these multiplications will be the mineral
r "modal" percentages by weight of the various minerals present,
nd their sum is the mineral or "modal" composition of the clay.

This process may strike one as very complex and laborious,
ut it is, in reality, very simple and rapid, particularly after one
as had a little practice. Let us take an example, a very simple
ne to begin with. A clay has the following chemical composi-
ion:

Per Mol.

cent. No.

Si<)_, 63 .36 = 1 .056

60

A 1,0,.

K,C) .

H,f) .

24.28

24.28

= 0.238

102

3 29

3 29

°°35

94

0 07

9.07

0 . 504

100.00

18

1 1 i JOURNAL OF THE

To calculate the orthoclase first, we multiply its molecular
weight, 556, by the unit molecular amount present, 0.035; U(
have then, 556 X 0.035 = l94& = <)r- The amount of AJ20|
required for orthoclase, 0.035, *s subtracted from the total amount ,

0.238, and the remainder, 0.203, serves as a measure for the
kaolinite. So we multiply the molecular weight of kaolinite,
258, by the unit molecular amount: 258 X 0.203 = 52-37 = Kl,
the percentage of kaolinite. All the K2() and the A 1 ' ): have thus
been assigned; the only mineral present now remaining to be
calculated is quartz. We therefore subtract from the total
molecular amount of SK)2, 1.056, the amount used for ortho-
clase (6 X 0.035 — 0.210) and that used for kaolinite (2 X 0.203 =
0.406), together 0.616; and multiply the remainder, 0.440, by the
molecular weight of quartz, 60. We thus have, 60 X 0.440 =
26.40 = O, as the percentage of quartz. We have now only to
consider the water. From the total molecular amount, 0.504,
we subtract that which was used for kaolinite, 2 X 0.203 = 0.406,
and multiply the remainder by the molecular weight of water,
18; the product is the percentage of moisture; thus, 18 X 0.098 =
1.76 = Aq. The mineral composition of the clay is, therefore, as
follows :

Kl (kaolinite) 5237

Or (orthoclase) 19 .46

Q (quartz) 26 .40

Aq (moisture) 1.76

99 99

This represents the actual mineral composition of such a simple
mixture with great exactness, provided of course, the chemical
analysis on which it is based is correct; and such a mode is fai
more accurate and trustworthy than the results of any "rational'
analysis made by the often recommended methods.

Let us take a more complex case, in which lime, soda, anc
ferric oxide are present in addition to the oxides present in tb
clay just calculated. Suppose that the composition by chemica
analysis is :

AMERICAN CERAMIC SOCIETY. 415

Per cent.

MoL

No.

63-36

I 056

23-76

O.233

O.52

O.OO3

O.69

O OI2

O.97

O.O16

I 63

O.OI7

907

O.504

IOO.OO

Si02..
A1203-
Fe203-
CaO..
Na20.
K20..
H20..

[ere (assuming the absence of calcite) the lime (CaO) forms
art of residual anorthite, CaO.Al203.2Si02, or possibly one
f the calcic zeolites, which contain hydrated anorthite mole-
ules. The soda (Na20) forms part of residual albite, Na20.-
i.l203.6Si02, or possibly may form part of a sodic zeolite, in which
ydrated albite or nephelite molecules are present,1 the difference
1 silica content of which is of negligible moment. The ferric
xide may be assumed to be present as limonite (Lm), Fe203.H20,2
he most common of the hydrated ferric oxides. The calculation
; briefly done thus: after calculating the orthoclase, albite, and
northite, the remaining alumina serves as a basis for calculating
he kaolinite ; the ferric oxide is the basis for calculating limonite ;
he silica remaining from the orthoclase, albite, anorthite, and
aolinite is the quartz, and the water remaining from kaolinite
nd limonite is moisture.

The slight deficiency in the summation is caused by the errors
1 not carrying out the calculations of the molecular numbers
eyond the third decimal place, which in this case all tend to low
esults.

The calculation is represented thus:

1 The presence of zeolites in clays, and their composition, cannot be dis-
ussed here; the lime and soda are present to so small an amount in usable
lays that errors in the assumptions above are of negligible consequence.

2 This is the formula for limonite as determined by E. Posnjak in an
ivestigation of the hydrated oxides of iron carried out in the Carnegie
eophysical Laboratory, and to be published in the American Journal of

416 JOURNAL OF THE

Or

= 556 X 0017

= 9-45

Al)

= 524 X 0016

= 8.33

An

= 278 X O.OI2

= 3-34

Kl

= 258 X 0.188'

= 48.50

I.in

= 178 X 0.003

= 0.53

Q

= 60 X 0.4582

= 27. 4»

Aq

= 18 X 0.125'

= 2 .25

99.88

If carbon dioxide (mol. wt. = 44) is found, it is assumed to be
present in ealcite (mol. wt. = 100); an equivalent amount of CaO
is allotted to the CO2 for ealcite, the ealcite is then reckoned by
multiplying the molecular number of C02 by 100, and the rest
of the lime is calculated to anorthite, as above.

Magnesia, MgO (mol. wt. = 40), may be assumed to be present in
chlorite, a complex mineral, or rather mineral group, into which
biotite usually alters; but it is best treated on the assumption
that it enters serpentine, 3MgO.2SiO2.2H2O (mol. wt. = 276), the
unit of calculation being one-third of the molecular number of
the MgO present. Any error involved will be quite negligible.

The micas remain to be considered. Of these, biotite may
be disregarded, partly because it is seldom present in more than
extremely small amounts in usable clays, and also because, if
it is desired to take it into consideration, its amount may, without
any practically appreciable error, be assumed to be that of
serpentine.

Muscovite is more commonly present, though in small amount.
Its composition may be considered as a mixture of orthoclase,
alumina and water, so that the percentage amount in a clay
cannot be calculated from a chemical analysis alone if orthoclase
and kaolinite are present also. However, as orthoclase anc
muscovite have about the same fluxing effect,4 it will serve all
practical purposes to calculate the potash of any muscovit(:

1 O.188 = O.233 — (0.OI7 + O.O16 rf- O.OI2).

2 0.458 = 1.056 — [(6 X 0.017) +,(6 X 0.016) + (2 X 0.012) 4
(2 X 0.188)].

3 0.125 = 0.504 — [(2 X 0.188) + 0.003].

4 Ries, "Clays," p. 102.

AMERICAN CERAMIC SOCIETY. 417

resent as orthoelase and the alumina as kaolinite, some of the
xcess silica being taken for this.

It will be obvious that the separate determination of potash
nd soda is necessary for the proper calculation of the feldspars,
n many analyses of clays, these are not separately determined,
ut are reported as "alkalies," that is, as the oxides. In the
ourse of analysis the alkalies are weighed together as chlorides,
nd to calculate these as oxides either the potash and the soda
lust be determined separately or the chlorides are assumed to be
lade up in arbitrary (and unknown) proportions of potash and
3da. The latter procedure is highly unscientific and unsatis-
ictory, and it should not be countenanced by workers . in
eramics.1

In order to show the actual results of such calculations as com-
ared with the compositions deduced from "rational" analyses,
tie calculated norms of several analyses of clays tabulated by
Lies,2 who also gives their "rational" analyses, are presented in
able I. As, in these analyses, potash and soda have not been
etermined separately, the calculations have been made on the
wo assumptions: (1) that the "alkali" is all potash, the alkali
ildspar being orthoelase; (2) that it is all soda, the alkali feldspar
eing then albite. The small amounts of magnesia are dis-
carded in these calculations.

Several points are clearly evident from a study of these calcu-
ited modes as compared with the "rational" analyses; and it is
) be remembered that these are only a few of those that have
een calculated. It is to be assumed that the chemical analyses
re fairly correct, in view of the simplicity of the determinations
f the few constituents present.

The first point is that the "rational" analyses do not generally
gree with the mode; indeed, some of them are hopelessly and
Imost ludicrously incongruous. In I, the R. A. ("rational" analysis)
hile showing about the correct amount of kaolin, gives several
mes the amount of "feldspar" that could possibly be present
n the basis of the alkalies. On the other hand, the R. A. of II

1 Cf. Washington, Prof. Paper, 99, p. 15; Mellor, "Quant. Anal.," p. 222.

2 H. Ries, "Clays," 1914, p. 67.

418

JOURNAL OF 'I'llF.

vo a

? t

w £

PQ a)

<! C

H <5

a

Z

0

0

0

o

C

0

<■<

VO

_

Ol

o

e> I

00

^j

f>

iO

er>

-i-

C

w

*"*

o

-

u

fa

CO

O

o

■/■

-1-

"" 1

Ov

r-

O

00

f*5

lO

rt

M

"2

0

Z

o

B

0

0

t^

Sd

O

O

O

o

o

O

o

t^-

01

»t

rt

o

r-»

vn

<*

CO

d

■*

00

il

=

i

&

o

vO
-4-

vO

(N

g

o

r*

8

«*>

X

(S

00

vO

QV

rO

n

fO

*■

iO

e

<*">

0

o

0

X

o

u*J vo

r)- - 0 H «*>

u-> tj- go r-. 0> "5 O

— .oonO C C -r ^

OOoOrOO • O 0 O vO

o o o

o o o o

O vO 00 00 00 tO

■J b O "• O flOO

ph 8

**

0

o

o

M

to

vO

*J-

c

ON

m

o

IO

•*

o

c

3

rO

O 0

0

C)

o

0

o

O

>o

-

0

0

o

o

o

o

r~

Ov

Tj"

vC

>o

o

o

o

vO

cs

C3

00

r—

S" c

11 jj
Pl, o

PO

IC

o

O

o

0

Ov

vO

(S

<N

^

vO

N

w

t^

•o

Tj"

iO

o

© 0

o

M

0

o

0

o

w>

H

O

0

o

o

0

O

(N

r^

M

O

IO

**

r^

£> C

IO

IT}

Ov

M

vO

o

cs

PU 1/

vO

in

o

o

o

M

O

O

c

r^

o

o

vO

O

-t-

c

O

Ov

0 0

o

w

O

o

o

o

^-

S£

o

O

o

o

o

O

o

•*

t^

crj

o

1> c
p< OJ

-*

<o

o

IO

Ov

00

cs

VO

M

0

o

O

00

CJ

vo

N

< J

0 r^

O t-

o

0

« o o o o s

cH <tj to g O <

AMERICAN CERAMIC SOCIETY.

419

0

O

O

0>

q

O

O

0

0
0

•<J-

0 vo
"3 00

u a

ss5

J-4

c ^

41 --

S

5 S

3 bO

<H

OJ O

■+

UO

P"5

M S

ro

O*

r~

>:3

00

00

<s

cd «

00

_c

— * c

s

0

<u *3

Ctf

h °s

W

ft M

lH

£

><i X

0)

0

CN

0

r-«

>

PO

0

"->

3

M

4-

M'

CO

•H

w

>>

JS

"3

rO 10

0)

O* CI

IS

6 M

£

u

§ S

5 o .3
•5 h4 W

3 3 _a

U CO cZi

420 JOURNAL OF THE

yields only o.io of feldspar although r.04 per cent of alkalies an
present, which demand either about 6 per cent of orthoclase <>r <
of albite. The presence of 3.34 of anorthite is also to be con
sidered. Nos. Ill, VII, VIII, and IX show the same feature
It cannot be objected that the difference is due to adsorbei
alkali — since the amount of this would be quite insufficient t
account for the discrepancy. No. X shows in its R. A. somewha
high feldspar, but rather of the correct order of magnitude, whil
the quartz is somewhat too low, and the kaolin too high. It i
evident from these results that there is no constant direction c
error in the figures of a rational analysis as regards any of th
constituents, and that they are therefore quite unreliable and un
worthy of use.

The second point brought out is that, in most of the R. A.'s
a very considerable proportion of the feldspar has been dissolve*
by the treatment with acid and has been reckoned as "Cla;
Substance." This is clear in II, III, VII, VIII, and IX.

It would also appear that the feldspar is generally almos
purely potassic, at least when little lime is present, though i
is also evident that the anorthite molecule is present in rathe
notable amounts in the feldspars of some clays. It is to b
remembered that the isomorphous presence of anorthite in th<
plagioclases increases their solubility in acids. The fact tha
the feldspars in the clays represented by the above analyses ar
chiefly orthoclase is indicated in the modal analyses VIII an'
IX, which show a deficiency in silica. This deficiency is cause
by the lesser molecular weight of soda, which takes more silic
to form albite than does potash to form orthoclase.

It need only be said, in conclusion, that the utter unreliabilit
of the rational analysis may be considered to be establishec
and that the term "rational," as applied to this method of d<
termining the mineral composition or mode, is not only a mi;
nomer but contrary to truth.

Attention must also be again called to the fact that, while tl
suggested method for calculating the mode may appear to 1
complex and "theoretical," yet it really is simple and express
very accurately the actual mineral composition. This is so, I
cause of the extremely simple character of the mineral mixtur

AMERICAN CERAMIC SOCIETY. 421

e simplicity of the mineral molecules, the impossibility of the
trance of other mineral molecules into the problem to any
sturbing extent, and finally the absence of the necessity for
nsidering the molecules of the minerals present as made up of
her assumed ones, a difficulty that is met with in the calcula-
m of the modes of igneous rocks when such minerals as alu-
inous pyroxenes, amphiboles, and micas are present.
This method of calculation is not the only one that might be
opted for arriving at the same results. Thus, without reducing
e percentages to molecular numbers, the proportions of kaolin,
dspars, and quartz can be calculated from the percentages
rectly. Thus the percentage of orthoclase can be calculated
)m that of potash by multiplying this by 1 .085 for the alumina
d by 3.832 for the silica, and adding the three figures. The
lount of alumina so obtained is to be deducted from the per-
itage of alumina shown by the analysis, in order to get the basis
r kaolin, and the amount of silica is to be deducted from that
the analysis for the balance that goes to quartz and kaolin.
le several factors necessary are obtained by calculating the
tios of the different constituents of each mineral to the unit
nstituent. This will yield the same results as the method de-
■ibed above, but will obviously call for much more numerous
Iculations. If tables are used the calculations involved in the
oposed method are almost entirely done away with.

Geophysical Laboratory,
Carnegie Institution of Washington,
August, 1918.

THE ACTION OF ACETIC ACID SOLUTIONS OF

DIFFERENT STRENGTHS ON A SHEET STEEL

ENAMEL.

Iiv I.Hon | FROST, Cleveland, Ohio.

For commercial reasons it became desirable to further prov
that a 20 per cent acetic acid solution exerts the greatest corrosiv
action on the type of vitreous enamel generally used on kitchei
utensils.

It occurred to us that it might be worth while to place the dati
resulting from our experiments on record, especiallv since it bear
out so well the work of R. D. Landrum1 on the same subject

The procedure was to allow equal volumes of the variou
strengths of acid to act on the inside surface of small enamele<
dishes and to determine the loss in weight resulting from thi
action.

The dishes for this purpose were prepared by coating smal
basins with, first, a coat of ground and then a thin coat of whifc
enamel. Over these was applied a good coating of the glaz>
to be tested. The molecular formula of this glaze was as fol
lows :

0.885 Na20
0.085 K20
0.015 CaO
0.015 MnO

] 2 .291 Si02
0.263 AI2O3 i 0.452 B203
I 0.701 F2

The enamel frit was milled with 8 per cent of Vallender Clay an
1 per cent of magnesium sulphate. Each dish was marked so
to show the per cent of acid to be used in it and these markinj
were thoroughly burned in. Only perfect dishes were used so I]
to minimize any chance of error. The dishes were then careful,
cleaned, dried on a hot plate and cooled in a desiccator, aft
which they were accurately weighed. The required amounts
distilled water and of technically pure, acetic acid to make
cc. of solution in each case were added from burettes and t
1 Trans. Am. Ceram. Soc, 14, 489 (1912).

AMERICAN CERAMIC SOCIETY. 423

olutions were taken to dryness on a hot plate, the temperature of
rhich was maintained as uniform as possible. This temperature
ras such as to cause a slight boiling but not sufficient for any
>ss from spattering of the solution. Evaporation required about
5 minutes. As soon as the dishes were dry they were removed
■om the hot plate and, after cooling, were washed and scoured
ith a soft abrasive so as to remove all material loosened by the
ction but so as not to affect the enamel itself. After rinsing
ith distilled water, the dishes were again dried on the hot plate,
Doled in the desiccator and weighed.

An attempt was made to keep the conditions of testing as
anstant as possible. In spite of this care some of the results
lowed inconsistencies which warranted their being omitted,
hese inconsistencies were due mostly to chipped rims and pin
listers.

The strengths of the acetic acid solutions used and resulting
teses in weight are tabulated as follows:

Per cent Per cent Per cent

acid. Gr. loss. acid. Gr. loss. acid. Gr. loss.

I 0.0056 17 0.0140 50 0.0168

2 0.0108 18 0.0151 60 O.OH7

3 O.O127 20 0.0160 70 O.OO76

4 O.O140 21 O.0214 80 O.OO26

5 0.0111 24 o\ 0223 90 0.0013

7 O.Q088 27 0.0184 IO° O.0003

9 0.0127 3° 0.0132 ...

15 0.0132 40 0.0179

The results are better shown by the curve (Fig. 1). This curve
oes not check exactly with those presented by Landrum but does
urn a very similar variation. The fact that there is a marked
eflection, greatest at 30 per cent instead of at 25 per cent as
jported by Landrum, might be explained by our having used the
■knically pure instead of the chemically pure acid. There

no apparent explanation for the irregularity between 2 per cent
id 9 per cent but from the results of several duplicate trials we
re convinced that this was not entirely the result of error.

It is evident from this curve that a 4 per cent to 5 per cent
jy volume) solution of such an acetic acid as would probably be

424

J01 RNAL OF THE

used for testing at most cooking utensil plants is well up in strength
as regards its action on the enamel surface. This per cent acid
is also of about the same strength as the strongest vinegar —
which is probably the most destructive agent to which an enameled

II

E; »-lL

^ "-Tf

--^

/

i

2s . 1 V

s

% I'M

r

$ rt-i-

v /

c J J]

\/

*"

J 1

itt

lit

Tit

5 7 9 >S n,8 70Z

J, }? fi

r i

0 \

c f

0 )

^ fOO

Pc cent Acid

Fig.

cooking utensil is subjected in ordinary use. The high point from
20 per cent to 25 per cent proves satisfactorily the point under
investigation.

COMMUNICATED DISCUSSIONS.

J. B. Shaw: It seems to me that this work would have a little
more value if the glaze selected for test had been of such a com-
position as to be less easily attacked by the acid. The glaze used
is an alkaline boro-silicate, very high in fluorine, and therefore
very soluble in water. While it is similar in its composition to
enamels, it would make a very low grade enamel for even cooking
utensils and could not be considered for use on ware requiring any
special quality as regards acid resistance.

The data obtained seems to prove roughly the original assump-
tion that about 20 to 25 per cent acid is most active on this
enamel, but the method of making the test admits so many

AMERICAN CERAMIC SOCIETY. 425

>ossibilities of error that the value of the data is open to question.
?or instance :

(a) There is no doubt that the strength of the acid varies in
very case throughout the period of evaporation.

(b) The rate of evaporation influences the results very ma-
erially.

(c) Scouring the surface with abrasives would surely introduce
ome error.

(d) Loss in weight would be proportional to the surface ex-
tosed. The same volume of acid does not necessarily act upon
he same area of surface.

It is noted that care was taken to keep conditions constant,
»ut after all precautions have been taken it seems likely that the
tossible error would exceed the difference between the results
•btained with the 4 per cent and 24 per cent acid solutions.

No solvent except water can be maintained constant in a com-
>osition during the period of evaporation to dryness. Because
if this fact it is impossible to obtain uniformly consistent results
ti such tests.

I believe a more reliable method of testing solubility — based
ipon the subjection of a given surface to a solution of constant
trength at constant temperature for a short period of time —
ould be devised.

E. P. Poste: An interesting feature in connection with the
etion of acetic acid on enamels came to our attention in the case
f a closed unit operating on strong acetic acid at elevated tem-
peratures and under several pounds pressure. An enamel for-
nula, which had been in successful operation in the canning in-
ustries under all possible conditions of acidity, and which had also
een used on a large number of open units and stills for the stronger
cetic acid solutions in various pharmaceutical and chemical
lanufacturing operations, was specified for this job. The enamel
l question failed where the vapors had come in contact with the
namel — but was in perfect condition where it had only been in
ontact with the liquid. In other words, the action of the acid
apors at elevated temperatures and under pressure was much
lore severe than that of the liquid. Similar observations have

426 Journal OF THE

been made in connection with hydrochloric acid. The volatility
of these acids would seem to account for the phenomenon. It
has no direct bearing on the action of acetic acid on cooking ware
unless it be the possible action of vapors in a covered kettle. How-
ever, the lack of pressures under these circumstances would seem!
to indicate that the case is not parallel.

B. A. Rice: Mr. Frost has very nicely accomplished his pur-
pose and checked previous investigations. It is of interest to
note that he secures the same peculiar dip in his curve — just be-
yond the high point — as did Landrum.1 The fact that the per-
centage losses obtained by Mr. Frost are slightly higher is proba-
bly accounted for by the fact that he used technically pure in-
stead of chemically pure acid.

Without a doubt the author has chosen the most desirable
method of testing the acid resistance of his enamel, that is,
by testing the enamel after it is burned on the ware.
However, it would have been interesting to have tried checking
this curve by subjecting the ground frit to the test outlined by
Poste.2 Poste compared his acetic acid curve with those ob-
tained by Landrum,1 but since he was using an enamel of fairly
high acid resistance, the high points of the curve were not brought
out as clearly as they might have been had a softer enamel, such
as is used in cooking ware plants, been used. Should the curves
resulting from the testing of an enamel by these two methods be of
the same general nature — though the actual gravimetric values
would differ — it is quite evident that testing the ground frit
would be a much more simple laboratory method of controlling
the acid resistance of enamels, once the acid resistance of a standard
enamel had been established.

As pointed out by Staley in his discussion of Landrum's arti-
cle, the method used by the author is accompanied by sonv
very undesirable features: (i) The different rates of evapora
tion, (2) the variation in the strength of solution — due to evapora
tion, (3) the constantly changing surface area exposed to the actioi
of the acid. These might be eliminated by simply covering th<

1 Trans. Am. Ceram. Soc, 13, 494 (191 1).

2 Ibid., 18, 137 (1915)-

AMERICAN CERAMIC SOCIETY. 427

ish with a watch glass and heating at a constant temperature
;low boiling for a certain length of time. This would allow two
leans of obtaining results — (1) the loss in weight of the dish, (2)
le determination of the amount of dissolved material in the solu-
on. The objection to this method would be the possible effect
r the vapors on the enamel. This factor could be reduced to a
linimum by filling the dish as full as possible.

It would also be interesting to see the tabulated results of the
:tion of cold solutions on the same enamel after standing a cer-
lin length of time. Would a curve plotted from these results
1 similar to the one obtained with the hot acids?

It might be well to mention also that tartaric acid from cooking
■apes, as in making grape marmalade, has a much more severe
:tion on enamels than acetic acid.

The foregoing points will undoubtedly be covered thoroughly by
ie Sub-committee on Enamels of the Committee on Standards,
merican Ceramic Society, in its work on the standardization
' enameled ware. Some very useful and instructive data will
I available for publication by this Committee in the near future.

R. R. Danielson: I believe a statement of the number of
ials used in each case would throw some light on the irregulari-
es in the data presented by the author. A limited number of
ials would undoubtedly cause irregularities — due possibly to
ich mechanical defects in the preparation of the samples as over-
jrning, etc., which might not be apparent under casual observa-
on.

I would suggest tests in the interval between 18 and 30 per
:nt and with variations of 1 per cent. Having this data, it is
assible that the author's data might approach more closely the
suits secured by Landrum and would at the same time make his
[suits more conclusive.

L. J. Frost: Mr. Shaw brings up some very interesting
[>ints. It is true that the glaze used was somewhat less acid
Csistant than the ordinary cooking utensil enamel, but it was
Cosen purposely on that occount so that the variations in the
Irve would be contrasted the more strongly.

In view of the fact that the same volume of solution was used

428 AMERICAN CERAMIC SOCIETY.

in dishes having the same dimensions and the same number of
coats of enamel in each case, and that the rate of evaporation
was practically the same with every dish tested, I feel that the
amount of error introduced here was too small to seriously alter
the results. Admitting that the strength of the acid solution
does vary in every case, throughout the period of evaporation, it
still seems that the action of the acid solutions of different strengths
should have the same relation as long as the rate of evaporation
does remain practically constant.

A simple method of testing, and one which would give more
accurate results in commercial investigations of this nature,
would be welcomed by the writer.

The suggestions of Mr. Rice that the curve given be checked
by the use of cold solutions, only, for a sufficient period of time,
and that this same enamel frit be also tested by the method used
by Poste, are very good. The objection to the latter method
for control purposes, however, is due to the fact that conditions
in sheet steel enameling often necessitate mill additions of so
varied a nature as to materially affect the acid resistance of the
completed enamel. If the frit alone is tested, this possible
variation is lost sight of.

The proposed improvements on the above tests are very worthy
of investigation.

Mr. Danielson is correct in his belief that over-burning and
other unavoidable conditions of manufacture do cause some
irregularity in the results. In running a large number of tests,
always with a 20 per cent solution and on the same enamel — •
which was much more acid resistant than the one given above — I
have found 1 to 1.5 mg. loss to be the usual limits of varia-
tion. However, when testing two dishes coated from the same |
lot of enamel and burned at the same time, the results have
checked exactly.

NOTE ON COLORS PRODUCED BY THE USE OF
SOLUBLE METALLIC SALTS.

Bv Paul G. Larkin, Denver, Colorado.

Introduction.

For a number of years soluble metallic salts have been used
n under- glaze color work. In 1910 experimental work developed
hat this method could be employed as a means of producing
)ver-glaze color in polychrome terra cotta.

The usual methods of production of polychrome terra cotta
nvolve the painting of colored glazes in the required areas with
1 camel's hair brush — or the use of shields or stencils to protect
:ertain areas while a colored glaze is sprayed on other portions
>f the piece. These, and numerous other devices, are objection -
.ble on account of the time consumed, the results obtained, and
he uncertainty of the whole process. They are presumably
00 well known to require further discussion here.

The use of soluble metallic salts offers a partial solution of this
>roblem. By this method the pieces requiring the specified
:olors are first sprayed with the glaze or slip to be used as a base —
he soluble salt solution then being applied to the indicated
ireas by hand painting. A steady hand and a little judgment
ire the only requisites for the application. The simplicity of
he operation and the rapidity with which the over-glaze solution
s applied are decidedly in favor of this method — when compared
vith the usual operation. To facilitate the application, it is
lecessary to have a base glaze or slip with an even, smooth sur-
ace such as would be obtained at the spray table from a water-
;lazed piece. A rough or granular-surfaced piece is painted with
lifficulty and seldom permits of smooth, clean-cut lines. The
jlaze used should contain a certain amount of gum arabic, molasses,
>r a similar substance affording a bond of sufficient strength to
nable the piece to withstand the handling required during the
olor application. Generally, standard finish terra cotta is cov-
red with an engobe having a clay content sufficiently high to

430 JOURNAL OF THE

give a bond which, with reasonable care, will undergo handling
without chipping or rubbing off.

For the past 2V2 years this scheme has been used at our plant
with good results. It has been used with matt, full-glaze, and
standard finish terra cotta. The application on matt glazes,
as a whole, gives the best results and the greatest color range —
producing colors of cleaner shade than those possessed by either
the full-glazed or standard finish ware.

The process is best adapted to color designs in which the area
to be treated is comparatively small — particularly if the solu-
tion is to be applied by a brush. Uniformity of color may be
obtained over large areas by spraying the solution, but this is
successful only in the case of flat surfaced pieces — where it would
probably be as economical to supply the required color by a glaze
containing metallic oxides.

In one instance, a color scheme required twelve colors. The
smaller color areas, ten in number, were obtained through the use of
soluble salts, while the larger remaining color areas were produced
by spraying colored glazes. The colors produced are not entirely
free from variation and in a large unbroken area are prone to
show unevenness. Nevertheless, the color is generally applied
to ornamental pieces and the lights and shadows offset the varia-
tions. It is on this last-named type of work that the application
of colored glazes gives the greatest difficulty, and, despite the
variation obtained through the use of soluble salts, the method
produces better results in a shorter time. Ornamental insert
pieces, in one, two, or more colors and commercial features, such
as name panels and trade marks in color, are easily and quickly
manufactured.

Application.

In the application of these color solutions, it is not advisable
to use a full brush of the solution as it spreads rapidly and, if
applied in this fashion at a point where a clean line is required,
gives a blurred outline. The better practice is to use a full brush
in the center of a color area and, as the supply of the solution in
the brush is depleted, to work smoothly and without undue
spreading to the outer edges in order to secure a cleaner line.

AMERICAN CERAMIC SOCIETY. 431

)ne application is sufficient if the surface is thoroughly covered.
Solutions that produce a streaked appearance after firing are not
mproved by a second coating applied in a direction at right angles
o the first application. The effect is merely changed from a
triped to a checkered one.

Preparation.

The following represents the method employed in the prepara-
ion of the different blends:

We arbitrarily adopted a standard solvent which consisted of
5 per cent water (not distilled) and 25 per cent glycerine
by volume). Standard solutions of nine soluble metallic salts in
his standard solvent were made by adding an arbitrarily adopted
weight of each salt to the solvent. A supply of these solutions
iras kept on hand, in glass-stoppered bottles.

To each standard solution was added 10 cc. of wood alcohol
o every 100 cc. solution (or roughly 10 per cent of wood alcohol),
["he different standard solutions were blended by the use of
mrettes on a volumetric basis. Aniline dye in solution in alco-
10I was added to the blends in the same proportion as added to
he standard solutions, i. e., 10 cc. aniline-alcohol solution to
ach 100 cc. of blended solution. For this purpose red, green,
>lue, violet, yellow, brown and black aniline dyes were used in
>rder to lend color to the solutions and thus aid in the applica-
ion. This method offers a simple basis for experimental work
>ut does not admit of ready comparison.

FORMULAS— Glazes and Engobe.

The following formulas represent the two types of glazes on
vhich the soluble over-glaze colors were applied:

Full Glaze (Fritted) (Cones 2-3).

0.25 K20
o 25 PbO
0.30 CaO
0.20 ZnO

0.325 A1203

2 . 20 SiO*

+ 6 per cent S11O2

+ 0.2 per cent uranium oxide

i ;•

JOURNAL OF THE

Matt Glaze (Raw Type) (Conks 2-3)

0 060 ECNaO
0.073 K20
0.032 Na20
0.004 MgO
0.413 CaO
0.418 ZnO

0.25 AI2O3

1 7 Si< )■•

The engobe used as the base for the standard finish terra cotta
samples had a composition as follows:

Standard Finish Engobe (Cones 2-3).

Batch weight.
Per cent.

28 .0 Feldspar

23 .0 Flint

1 0.0 Cornish stone

4-5 F"t
21 .0 Ball clay
12.0 Georgia kaolin

1 .5 Borax

Frit.
Per cent.

75 .0 Red lead

25 .0 Flint

100. 0

(Melted down in a frit kiln.)

Standard Solutions.

The standard solutions used were prepared as follows from
technically pure salts:

No. 1. 100 cc. solvent.

50 g. ferric nitrate.

10 per cent wood alcohol (by volume).

No. 2. 100 cc. solvent.

50 g. cobalt chloride.

10 per cent wood alcohol.

No. 3. 100 cc. solvent.

50 g. uranium nitrate.
10 per cent wood alcohol.

No. 4. 100 cc. solvent.

50 g. manganese nitrate.
10 per cent wood alcohol.

AMERICAN CERAMIC SOCIETY.

433

No. 5.1 100 cc. solvent.

25 g. chromium nitrate.
2 g. zinc chloride.

10 per cent wood alcohol (by volume).
No. 6. 100 cc. solvent.

50 g. nickel nitrate.

10 per cent wood alcohol.
No. 7. 100 cc. solvent.

50 g. copper nitrate.

10 per cent wood alcohol.
No. 8. 100 cc. solvent.

50 g. tin chloride.

10 per cent wood alcohol.
No. 9. 100 cc. solvent.

50 g. zinc chloride.

10 per cent wood alcohol.

1 We attempted to use larger amounts of the chromium nitrate in No. 5
it had difficulty in getting it into solution. The difficulty was due to the
iterioration of the salt. Zinc chloride was used because a salmon color
is desired.

Blends for Green Colors on Zinc Bearing Glazes.

Solution.
?erric nitrate ....
Cobalt chloride . .
Jranium nitrate .
Nickel nitrate . . .
Copper nitrate. . .
Manganese nitrate
Solvent

No. O

1

25

25

10 16

20 20
10 5
10

9

2o Zc

4 3

5 6

6C 6U 6° 6° 6U

4-5 2.5 8.0 5

17.0 3.0 2.5 2.0 4

19.0 12.5 10. o 10. o 8 8 8

4.5 9.5 10. o 5.0 8 8 8

50

25 25 25 25 25 25

To each blend, 5 cc. of aniline dye in alcohol solution was
ided.

Each individual blend in the above table produced a green on
latt glazes.

Blends Nos. G-4 and G-5 were the best greens obtained and
ere blue-green on matt glazes and blue on full or bright glazes
ith a slight green cast. Blue on standard finish ware.

Attempts were made to produce a green, using a combination
i solutions No. 7 (copper nitrate), No. 8 (tin chloride), and No. 9
zinc chloride). The results obtained on matt and full glazes

, 34 .!' HJRNAL OF THE

were not satisfactory. The color in nearly every ease varied
greatly blotched in one part with a dark metallic green-black
and a light dirty green in another part. Presumably, the base
glaze used was not adapted to the production of a copper-green.
Incidentally, green colors were obtained through the use of
uranium, but were not stable or dependable and were difficult
to combine in harmony with other colors.

BLUE COLORS.

The lighter blues were produced by blending the standard solu-
tion of cobalt (solution No. 2) with standard solutions of man-
ganese (No. 4), nickel (No. 6), and iron (No. 1), in varying propor-
tions and in some cases by diluting the cobalt solution.

Blue No. 1 :

7.5 cc. solution No. 6 (nickel nitrate).

2.5 cc. solution No. 1 (iron nitrate).
15 .0 cc. solution No. 2 (cobalt chloride).
25 .0 cc. solvent.

5 .0 cc. aniline dye in alcohol.
Blue No. 2 :

10 cc. solution No. 2 (cobalt chloride).
40 cc. solution No. 4 (manganese nitrate).

5 cc. aniline dye in alcohol.
Blue No. 3:

15 cc. solution No. 2 (cobalt chloride).
15 cc. solution No. 4 (manganese nitrate).

5 cc. solution No. 6 (nickel nitrate).
10 cc. solution No. 8 (tin chloride).

5 cc. solvent.

5 cc. aniline dye in alcohol.

Darker blues were produced by increasing the concentration
of solution No. 2 (cobalt) by the further addition of cobalt chlo-
ride. A good dark blue had the following composition:

Blue No. 4:

40 cc. solution No. 2 (cobalt chloride).
8 cc. solution No. 4 (manganese nitrate).
2 cc. solution No. 1 (ferric nitrate).
5 g. cobalt chloride to blend.
5 cc. aniline dye in alcohol solution.

AMERICAN CERAMIC SOCIETY. 435

BROWN COLORS.

Standard solutions No. 1 (ferric nitrate), No. 3 (uranium
nitrate), No. 4 (manganese nitrate), No. 6 (nickel nitrate),
and No. 5 (chromium nitrate) were blended to produce brown
colors. In general, the efforts were not successful — the colors
being light and of a disagreeable shade. The best colors ob-
tained consisted of blends as follows:

Brown No. i :

19.0 cc. solution No. 4 (manganese nitrate).
7.5 cc. solution No. 1 (ferric nitrate).

1 1 .5 cc. solution No. 6 (nickel nitrate).

12 .0 cc. solution No. 5 (chromium nitrate).
5.0 cc. aniline dye in alcohol.

Brown No. 2 :

10 cc. solution No. 5 (chromium nitrate).
18 cc. solution No. 4 (manganese nitrate).
2 cc. solution No. 3 (uranium nitrate).
5 cc. aniline dye in solution.
Brown No. 3 :

30 cc. solution No. 1 (ferric nitrate).
10 cc. solution No. 6 (nickel nitrate).
10 cc. solution No. 4 (manganese nitrate).
5 cc. aniline dye in solution.

GRAY COLORS.

Several acceptable gray colors were produced in blends of solu-
tions No. 1 (ferric nitrate), with a small quantity of solution
No. 2 (cobalt chloride), No. 4 (manganese nitrate), and No. 6
(nickel nitrate). A tendency towards brown was noted in the
application on full glazes or enamels. The variation is also
marked.

Gray No. 1 :

0.5 cc. No. 2 (cobalt chloride).
24.5 cc. No. 6 (nickel nitrate).
25 .0 cc. No. 4 (manganese nitrate).

5 .0 cc. aniline dye in alcohol.
Gray No. 3 :

25 .0 cc. No. 1 (ferric nitrate).

1 .0 cc. No. 2 (cobalt chloride).
10. o cc. No. 4 (manganese nitrate).
14.0 cc. No. 6 (nickel nitrate).

5 .0 cc. aniline dye in alcohol.

4,^> JOURNAL OF THE

YELLOW AND ORANGE COLORS.

The following blends represent the best colors obtained in com-
bining solution No. i (ferric nitrate), solution No. 3 (uranium
nitrate), solution No. 4 (manganese nitrate), and solution No. 5
(chromium nitrate). The colors are the results of experimental
work in an attempt to produce yellow and orange colors

Y-i:

20 cc. No. 3 (uranium nitrate).
20 cc. No. 4 (manganese nitrate).
10 cc. No. 5 (chromium nitrate).
5 cc. aniline dye in alcohol.

Y-2:

10 cc. No. 1 solution (ferric nitrate).

9 cc. No. 3 solution (uranium nitrate).
16 cc. No. 4 solution (manganese nitrate)
15 cc. No. 5 solution (chromium nitrate).

5 cc. aniline dye in alcohol solution.

Y-3:

20 cc. No. 3 solution (uranium nitrate).
10 cc. No. 4 solution (manganese nitrate)
20 cc. No. 5 solution (chromium nitrate).
5 cc. aniline dye in alcohol solution.

-*■

5 cc. No. 1 solution (ferric nitrate).
20 cc. No. 3 solution (uranium nitrate).
20 cc. No. 5 solution (chromium nitrate).

5 cc. solvent.

5 cc. aniline dye in alcohol solution.

In attempting to introduce tin chloride as another component
in Y-3, we obtained a brown -yellow of unpleasing shade.

The combination of an insoluble stain and a soluble salt gave
better results in the production of yellow than the use of soluble
salts alone.1

The following stain was mixed dry, calcined to cones 2-3, wet
ground and sieved through a 150-mesh screen:

1 Trans. Am. Ceram. Soc, 19, 653 (1917).

AMERICAN CERAMIC SOCIETY. 437

Yellow Stain.

White lead 22.6

Antimony oxide 20.2

China clay 15.8

Tin oxide 14.4

Calcium carbonate 1 1 . 2

Sodium nitrate 15.8

After calcination at cones 2-3, the stain was dense, brown-
ellow in color and rather hard.

This stain, suspended in standard solution No. 3 (uranium ni-
rate), produced a good orange-yellow color. The blend had the
Dllowing composition:

Y-8:

35 cc. solution No. 3 (uranium nitrate).
• 30 g. yellow stain (150 mesh).

35 cc. solvent.

(No aniline dye required.)

Brush application shows variation. The color is good. Further
ttempts to improve the color by the introduction of 150-mesh
utile produced a lighter color with a brownish cast — -not equal

0 Y-8 in strength. The Y-8 yellow is not as good on full glaze
nd on standard finish produces a faint yellow with no bond
•xisting between the stain and the engobe:

Comparative Costs.

Some information may be gained on the cost of application
>f these soluble colors from the following data on a standard
inish corner stone and a name panel:

A 40" X 20" rectangular corner stone with 28 Y counter-sunk
etters required 61 •> hours labor at $0.20 or $1 .30 and materials
osting $0.05, making an approximate total of Si .35.

The standard finish name panel contained 26 double Y-eounter-
unk letters with an approximate height of 8". Two colors were
tpplied, one to the outline V and a contrasting color to the deep

1 or body of the letter. The labor necessary to complete the

438 AMERICAN CERAMIC SOCIETY.

Color application cost $1.00 and the materials added $0. 22,

making an approximate cost of $i .82.

Conclusion.

Naturally, there are certain colors whose pyro-chemical re-
quirements cannot be satisfied by a solution application, but a
far as this scheme can be carried out, it solves some very per
plexing questions in polychrome terra cotta and doubtless can
be employed in other fields.

As noted before, it is labor saving; it gives good results; clean
lines are a matter of manipulation — except possibly on a full
glaze where the color "creeps" during the firing operation (allow-
ance should be made for this movement at the outside edges of
the color area).

The addition of a lead frit or flux to a standard-finish engobe — ■
making proper allowance for this addition so as not to bring the
engobe to a glaze or semi-glaze condition — will improve to a
marked degree the colors produced by the use of soluble salts on
standard finish ware.

Denver Terra Cotta Company,
Denver, Colorado.

NOTE ON THE SINTERING OF MAGNESIA.

By John B. Ferguson.

Prior to the war, pyrometer tubes, furnace tubes, and similar
are made from magnesia were imported from Germany. This
mrce of supply is now cut off and a new source has not yet ap-
eared. The sintering of chemically pure magnesia has been
snerally regarded as difficult, if not impossible, and the refractory
rticles referred to are, in fact, found to contain about two per
?nt. of forsterite (2MgO.Si02), together with about one per

Fig. i.

tnt. of other impurities (estimated roughly under the micro-
:ope), which act as a bond. It was thought of interest, there-
ire, to put upon record some observations on the sintering of
Iky, calcined, pure magnesia such as is used to insulate small elec-
ic furnaces.

In determining the melting points of cristobalite and tridymite,1

special electric resistance furnace of the cascade type was em-

loyed and the two heaters were insulated from each other by a

1 J. B. Ferguson and H. E. Merwin, Am. J. Sci., 46, 417 (1918).

440 JOURNAL OF THE

layer of silky, calcined, magnesia powder. The furnace was main
tained for some hours at temperatures ranging from 1600
to 17200 C and the magnesia packed so rapidly that it was neces-
sary to add more of it every hour. Upon dismantling the fur-
nace, the magnesia was found to have formed a dense cake sur-
rounding the inner heater, as shown in the accompanying photo-
graph. The cake had considerable mechanical strength. A
sintering of this character has not before been observed in our
furnaces, which operate usually below 16000 C.

A microscopic examination of this cake by Dr. H. E. Merwin,
of this laboratory, did not disclose any binding material; only
periclase (crystallized MgO) with less than one-half of one per
cent of forsterite (2MgO.Si02) could be identified. The crystals
of periclase were interwoven to a considerable extent.

The silica content of the cake was found by chemical analysis
to be o . 44 per cent.

The manufacture of magnesia ware of sufficient mechanical
strength without a binder would, therefore, appear feasible if
the product were fired at temperatures between 16000 C and 17000
C for some hours.

Geophysical Laboratory,
Carnegie Institute of Washington,
Washington, D. C.

:

AMERICAN CERAMIC SOCIETY.

441

AMERICAN CERAMIC SOCIETY.

Acquisition of New Members during September, 1918.

Associate.

litsu Kato,

too Fire Brick Co.,

too-ko,

kayamaken, Japan.

R. B. Gilmore, W. H. Clark,

East Liberty Y. M. C. A. Dow Chemical Co.
Pittsburgh, Pa. Midland, Mich.

>laf Anderson, F. H. Reagan,

tatsgeolog, Locke Insulator Co.,

lineralogisk Museum, Victor, N. Y.
jistiania, Norway.

. Zach Olsson,
30 Pearl St.,
lew York City.

J. S. Brogdon,
70^2 Peachtree St.,
Atlanta, Ga.

F. A. Harvey,
Solvay Process Co.,
Syracuse, N. Y.

Chas. C. Bacon,
Ross-Tacony Crucible
Co., Tacony, Philadel-
phia, Pa.

Contributing.
American Trona Corporation,
724 S. Spring Street,
Los Angeles, Cal.

442

JOURNAL OF THE

THE HONOR ROLL *

OF THE

AMERICAN CERAMIC SOCIETY

Howard C. Arnold
Kenneth B. Ayer
H. H. Bartells
Charles E. Bates
Harold A. Best
Clarence J. Brock bank
Wilbur F. Brown
L. R. Bucher
Louis L. Byers
John A. Chase
O. I. Chormann
Frederic C. Clayter
H. F. Crew
Richard D. Ferguson
Robert F. Ferguson
Kenneth I. Fulton
Chas. F. Geiger
Albert C. Gerber
Perry D. Helser
F. J. Hoehn
H. W. Holmes
Herford Hope
Roy A. Horning

Thomas N. Horsley
Walter L. Howat

F. Sumner Hunt
Chas. E. Jacquart
Joseph A. Martz
R. W. Miller
Earl J. Moore
Fred A. Morgan
James S. McCann
Edward Orton, Jr.
Walter S. Primley
C. E. Ramsden

G. W. Rathjeins
Guy L. Rixford
H. S. Robertson
Geo. M. Schaulin
Ira E. Sproat

F. L. Steinhoff
Chas. W. Thomas
E. H. Van Schoick
Francis W. Walker, Jr.
William G. Whitford
J. B. Yowell

* This list is not complete. Information in regard to those who are in
some branch of the service and whose names are not included in the above
list will be appreciated.

C. F. Binns, Secretary, Alfred, N. Y.

JOURNAL

OF THE

AMERICAN CERAMIC SOCIETY

A monthly journal devoted to the arts and sciences related to
be silicate industries.

ol. 1 July, 1918 No. 7

EDITORIALS.

LOOKING FORWARD.

With the practical ending of the war we look back with pride
t the achievements of the Society and its members in the solving
f the problems arising in its successful prosecution and in meet-
lg the difficulties arising in the normal conduct of the various
ranches of the industry. The ceramic industry has kept pace
rith the advancement of the other great industries during this
eriod of trial and we have seen the successful placing of several
iew and important products on a firm and lasting manufacturing
iasis in this country — notably the production of optical glass,
hemical glassware, chemical porcelain, etc.

Looking forward, we see some new and some old problems
fhich remain to be faced.

The scarcity and steadily rising cost of fuel of all kinds is
timulating interest on the part of our manufacturers in improved
ypes of kilns and methods of combustion. The suggestion that
t our annual meeting in February we have a session at which the
lerits of the various types of patented dryers, gas-producers,
unnel kilns and continuous kilns be discussed from a technical
nd economic standpoint, is an excellent one and it is to be hoped
hat an opportunity for a discussion of this kind will be afforded.

In this number of the Journal, Dr. Ries gives an excellent
esume of the available and prospective sources of high-grade
lays in the United States. Linked as it is with the question of
he production of high-grade pottery wares from all American

m JOURNAL OF THE

materials, the problem of the location of new deposits of Ameri
can kaolins and of improvement in the methods of their purifica-
tion should not cease to receive the attention of OUT geologists
and technologists. We arc faced with numerous other problems
awaiting attention and solution.

PROFESSIONAL DIVISIONS.

As a culmination of the discussion which has taken place of
late in reference to the formation of industrial divisions of the
American Ceramic Society, the Board of Trustees has recently
affirmed the following motion:

Professional divisions may be formed in the following ways:

i. "When the initiative in the formation of divisions is taken by members
who are interested, a petition may be presented by not less than ten members
in good standing, of whom three or more shall be Active, who are interested
in some phase of ceramic work sufficiently broad to warrant the formation
of a special division. This petition shall go to the President who shall then
appoint a representative committee to consider the advisability of forming
such a division, and to proceed with the organization if the decision is favor-
able.

2. When the initiative is taken by the Board of Trustees in order to stimu-
late the growth of the Society, the President shall appoint a representative
man to furnish the initiative. He shall select his own committee."

The steady growth and enlargement of the scope .of the ac-
tivities of the Society warrant this action by the Board at this
time. Our Society has outgrown the old idea of each member
having a broad interest in all kinds of ceramic work. If we are
going to maintain solidarity in technical ceramics, as it has been
in the past through the American Ceramic Society, we must fur-
nish opportunity for specialization. Professional divisions will
furnish this opportunity.

It is to be hoped that there will be no delay in the organiza-
tion of the respective professional divisions. The problem of
the formation of several divisions of the Society at once is not a
difficult one and is only a question of which branches of the in-
dustry offer the most fertile fields for stimulation and growth.
It would appear that the early organization of the glass and pot-
ten- divisions is particularly desirable.

AMERICAN CERAMIC SOCIETY. 445

In the past the importance of quickening the interest of a larger
roportion of the glass industry in the Society has been too little
cognized. The technical achievements and awakened interest

this highly important branch of the ceramic industry — through
le demands of the war for special glasses — should not be permitted
> die out.

It is needless to mention the importance of the active support

the pottery industry to the future welfare and advancement
r the Society and the necessity of stimulating research on the
'oblems arising in this basic industry. This is especially vital

view of the ending of the War and the consequent looming up
' the old and serious problem of again meeting the competition
j the foreign wares.

The organization of professional divisions for the other branches

the industry is of equal importance and in passing we may
astily mention the advisability of the formation of divisions on
lamels, refractories, abrasives, etc.

The securing of an adequate supply of professional papers for
le issuance of a Journal of even the present modest size has
roven somewhat difficult. Through the cooperation of the Pro-
ssional Divisions and Local Sections there should be an ade-
uate flow of varied technical contributions which will assure
le future success of the Society's publications.

ORIGINAL PAPERS AND DISCUSSIONS.

THE OCCURRENCE OF HIGH-GRADE AMERICAN

CLAYS, AND THE POSSIBILITIES OF THEIR

FURTHER DEVELOPMENT.1

Kv H. Hues.

Introduction.

Although the United vStates has always ranked foremost amonj
the countries of the world as regards its mineral resources, never-
theless, a not inconsiderable quantity of certain raw materials
has been obtained in the past from foreign sources. This was no
doubt due in part to the comparative ease and cheapness with
which some of these could be obtained, and also to the fact that
many of them were of high grade, the result being that domestic
consumers in many instances imported certain necessary mineral
products rather than search for them in this country.

With the beginning of the great war, certain mineral products,
hitherto obtained from central Europe, were almost immediately
shut off, while later, owing to the scarcity of ships, it became
necessary to further curtail the imports — even of those that could
still be imported from neutral or allied countries.

This curtailment of imports has in some respects been beneficial
to the United States, for it has stimulated the search for, and
development of, good deposits of useful minerals within our
boundaries, and it has, moreover, shown the possibility of using
low-grade domestic materials hitherto disregarded.

To quote one case outside of the field of clay. For many
years we have imported emery from Turkey and Greece. Since
1 By permission of the Director, U. S. Geological Survey.

AMERICAN CERAMIC SOCIETY. 447

2 war we have begun the development of a high-grade emery
Id in Virginia, which apparently contains a large tonnage of this
rasive.

While the clay imports form but a small percentage of the total
ports of mineral products, nevertheless, taken by themselves
;y make up no small tonnage, this in 1916 amounting to
8,866 short tons, valued at $1,504,233, and in 1917 to 239,318
>rt tons, valued at $1,442,059.

The clays which have been imported into the United States
^lude those from central Europe which were used in the manu-
:ture of glass pots, graphite crucibles, enamels, etc., and the
ina clays and ball clays of England, much sought after by
mufacturers of whiteware, electrical porcelain, paper, etc.
While it is true that the supply of the English clays has not
en completely shut oft, it has recently been diminished on
:ount of dock repairs at Fowey, and still further restricted at a
ry recent date (August), by an order of the British government —
entially prohibiting the exportation of clay in British bottoms
Boston, New York and Philadelphia.

With the shutting off, therefore, of so much material of value
the ceramic industry, it becomes imperative to determine
tat means can be, or have been taken in this country to replace
ported clays, or what the possibilities are for increasing the
tput of those high-grade ones now being obtained.
It is, of course, well known that kaolin, ball clays, paper clays,
d glass-pot clays have been mined in this country for some time,
t nevertheless, they have not fully met the domestic demand
two ways: firstly, there has been an insufficient quantity and
;ondly, many manufacturers have claimed that, for certain
rposes, it was desirable, if not necessary, to include some of the
eign clays in their mixtures — or in special cases to use them
rlusively. This, therefore, brings up two problems, viz.: first,
1 the supply of those clays already known and being mined be
•reased, and second, can we find additional supplies of the types
eady known in this country, or deposits of new types?
Fo accomplish this requires two distinct lines of investigation,
)logic and technologic.

448 JOURNAL OF THE

The geologic investigation serves to show the areas in vvhicli
new supplies may be looked for, their probable extent, size and,
therefore, tonnage. Thus, it saves the prospector and miner
much time and money in his search. It also serves to give in
formation regarding the character of the raw materials, their
visible impurities, etc.

The technologic study takes up the further development of the
problem and attempts to work the newly discovered materials
into different mixtures, or possibly find new uses for those deposits
of clay already exploited.

It is to the first of these that I shall give especial attention
touching the latter but briefly.

High-Grade Clay Deposits of the United States.

Any one familiar with our sources of raw materials used in the
manufacture of the high-grade products mentioned above, knows
that most, if not practically all, of them are located in the eastern
half of the United States, and while some of value may occur in the
far west, they have not been drawn upon by the factories of the
eastern and central states — due probably because of the long
hauls and consequently the high freight rates.

During the past summer and spring, the writer has had occasion
to visit most of the localities in the eastern half of the United
States which are capable of supplying raw materials of high
grade to the industries already referred to, and in the present
paper it is proposed to outline briefly what the chances are for
obtaining supplies and also to touch briefly on their present uses.'

The area under discussion embraces that portion of the United
States lying east of the Mississippi and in addition portions ol
Arkansas and Missouri.

For purposes of convenience of discussion, we may divide thi:
region into several geographic subdivisions as follows:

i. The Atlantic Coastal Plain belt, extending from New Jersey
to Florida and Mississippi inclusive.

2. The so-called Embayment Area, reaching from northen

Mississippi and Texas northward into southern Illinois and in

1 The detailed results of this work will appear in a bulletin of the U. i\

Geological Survey.

AMERICAN CERAMIC SOCIETY. 449

jding western Tennessee and Kentucky, southern Illinois,
utheastern Missouri, and eastern Arkansas.

3. The Piedmont plateau and mountain belt which lies just
;st of the Atlantic Coastal Plain belt and extends from Con-
cticut to northwestern Alabama, and westward to the Appala-
ian mountains.

4. The Coal Measures region of the central and southern states.

5. The Missouri flint-clay areas.

1. The Atlantic Coastal Plain Belt. — This region is underlain
■ a series of more or less unconsolidated sands and clays, which
p gently seaward — the most valuable clays having been found
nost exclusively in the Cretaceous formation, or oldest
le of the series, the only exceptions being certain localities in
sorgia and Florida. This means, therefore, that the high-
ade clays so necessary to many industries during the war are

be sought for mainly along or towards the inner side of this belt.

In tracing this formation from New York southward to Georgia

d westward to Mississippi, we find that it shows considerable

nation both as to the character of material and the possible

nnage.

The areas of clay in New York are small and of value chiefly

r fire brick, but the extension of this clay belt across New Jersey

well known; indeed, the products of the Woodbridge and
ighboring clay pits are so familiar to us as to need no further
2ntion. Continuing to the southward, the same belt crosses
aryland, but shows an entirely different character and is of
minished importance. It is true that along some of the rivers
wing into Chesapeake Bay, such as the Severn, white refractory
lys have been reported and do exist, but the deposits are lens
aped and, so far as discovered, are not of large tonnage. There
e also, in this same belt, some buff-burning clays of moderate
fractoriness which have varied uses in the metallurgical in-
istries, but the individual masses are small and indefinite,
le clay from some of these localities is considered adapted for
e in the crucible trade.

In Virginia, the Cretaceous for mation carries, so far as accessible,
1 high-grade clays, and the same is true in North Carolina,

I.S"

|< IURNAL l '!• THE

but then to the southward, in South Carolina and Georgia, tlii
same formation develops a scries of clay beds of striking charaete
and high commercial value. We refer to the deposits of whit
clays — variously known as kaolins and plastic kaolins. Tin
materials at the present time are widely used in the paper tradi
and to a lesser extent in the manufacture of whiteware, electric!
porcelain and wall tile.

These plastic kaolins are found in certain areas in Bibb, Twiggs
Wilkinson, Washington, Glasscock, Jefferson and Richmon<

Fig. i. — Removing sandy overburden from white clay with steam shovel
at Gordon, Georgia.

Counties of Georgia; and in Aiken, Edgefield, Lexington, am
Kershaw Counties of South Carolina. In Georgia, the clay i
worked chiefly in the Dry Branch and Mclntyre districts, while i
South Carolina, the developments have been chiefly in the Aike
district.

Those of Georgia are usually washed and yield a high pei
centage of washed product, while those of South Carolina are pt
on the market mostly in the un-washed but pulverized form.

AMERICAN CERAMIC SOCIETY.

451

n view of the widespread demand for these clays, a few words
y be said regarding them.

Ahese clays are always in the form of lenses, overlain usually
iron-stained sands. The lenses are of variable size, some of
m being large. The overburden of iron-stained sand varies
n 3 to 30 feet. The clays are worked at some 15 localities,
there are, no doubt, many other deposits not yet developed,
that the production could be expanded by this means. It

G. 2. — Excavating white clay with steam shovel at Mclntyre, Georgia.
The overburden is cross-bedded iron-stained sand.

Id be still further increased by working the washing plants
to their capacity. This latter would mean a production
bably 30 per cent greater than at present. But here we
ounter the question of labor supply and transportation
lities — both of which must be improved to aid in further
put.

Certain of the deposits in Georgia contain beds of bauxite
bauxitic clay associated with the white clay.

452

JOURNAL OF THE

The Cretaceous formation swings around to the westward and
is found in Alabama and northeastern Mississippi. There ;nc
white clay lenses here, possibly not as extensive as the Georgia
and South Carolina ones, and undeveloped. Moreover, they
occur in a higher geologic horizon than those of Georgia.

Those in Alabama are found chiefly near Marion, while those of
Mississippi are distributed around Iuka, Tishomingo County.
The clay lenses in these two states are probably not as large as the

•v

■ > . ■ .

7 F:%t

X?*J-£

Fig. 3. — White clay overlying bauxite in Sv/eetwater mine, near Ander-

sonville, Ga. Railroad tracks at upper limit of bauxite. White clay

overlain by iron-stained sands.

Georgia and South Carolina ones, nor are the materials always a
plastic.

East of the Georgia belt, there are some white clays in th
higher lying Tertiary formations, but they have not been utilizec
One of the best known overlies the bauxite in the Sweetwatf
bauxite pit, 5V2 miles east of Audersonville, Sumter Co.

There is also the well-known white plastic clay from Florid;

AMERICAN CERAMIC SOCIETY.

453

hich is worked in Putman and Lake Counties, at points about
o miles apart, but it probably underlies a considerable area in
etween and there is a possibility of much greater production.

2 . The Embayment Area. — The coastal-plain formations, already
:ferred to, swing northward from northwestern Mississippi and
>rm a belt extending northward up the Mississippi Valley into
mthern Illinois. The belt also extends for some miles to the
ist and west of the Mississippi River. It is so-called because

Fig. 4. — Dredging clay at Edgar, Florida.

he Cretaceous and Tertiary formations found here were de-
osited in a great northward reaching embayment of the sea,
nd it is here that some of our most important clays are found
id where there is, we believe, great possibilities for future de-
elopment.

Beginning at the south, we have in northern Mississippi several
nportant clay formations which deserve careful investigation,
ut we cannot speak definitely regarding all of them as yet.

454

JOURNAL OF THE

Tlu- hist of these areas lies in the western half of aorthenj

Mississippi the formation of Tertiary age, known as the Wilcox,
containing a great series of white, grayish white, and light pink
clays.

If we refer to a geologic section across this portion of the state,
we notice that the beds dip westward. Consequently, those
of the west half of the area will not be found in the eastern half
because they do not exist there-having been eroded — and con-

FiG. 5. — Clay pit in Wilcox formation near Holly Springs, Miss. To the
right of view the sand rises almost to the surface. At the left of
view the clay rises to the surface. This shows the irregular
occurrence of the clay in this region. Dark over-
burden is loam.

versely, those of the eastern half will be found only at consider
able depths in the western half.

We separate this area into two halves because the clays ar<
somewhat different. The western half yields clays of tougl
nature, high plasticity, buff-burning character, dense burninj
nature, and often high transverse strength. They vary in thei

AMERICAN CERAMIC SOCIETY.

455

efractoriness. Some are dark colored — due to the presence of
inely divided coaly material. The better ones are of the types
;hat can be used in glass-pot mixtures, graphite crucibles, and
n the bond of abrasive wheels. Others can be employed in stone-
ware, etc.

The clays of the eastern half occur in lenses, which sometimes
aass rather rapidly into sand beds. The clays themselves
:requently show thin layers of sand. They are now used only

Fig. 6.-

-Pit of refractory bond clay ten miles west of Enid, Miss,
overburden is loam.

The

for common stoneware, but their value after washing deserves
to be considered. The silica content is high, and the iron oxide
in published analyses ranges from 1.5 to 4.5 per cent. We ques-
tion, however, whether we have here a grade of clay as good as
those in the western half of the area.

The Tallahatchie county clays, although known for two or
:hree years, are not yet fully developed and we believe that we
lave here an important clay region for the crucible and glass pot

456

JOURNAL OF THE

trade, the demand for which is increasing steadily. The output
can be greatly increased by better mining and by better roads and
transportation facilities. It is curious that these- clays should
have remained undeveloped so long, but this may be due to the
lack of outcrops, for the clays are usually covered by a sandy
loam of varying thickness. However, their white outcrops in the
bluffs bordering the Mississippi River bottoms should have
attracted attention long ago.

Fig. 7. — Erosion gullies near Huntington, Tenn., showing natural exposure
of the light-colored Lagrange clay.

Proceeding northward from Mississippi, through Tennessee
and Kentucky, we have the well-known clay belt which has
supplied us with ball clays and other grades. The main forma-
tion of this belt, known as the Lagrange division of the Tertiary, is
really an extension of the Wilcox formation of northwestern
Mississippi, but the change in the character of the contained
clays is as great as that noted from New Jersey to Maryland.
Only in the Southern part of Tennessee, as around Lagrange,

AMERICAN CERAMIC SOCIETY.

457

do we find clays approaching the northern Mississippi ones in
character. Moreover, within the Lagrange formation, the best
clays have been found near its eastern boundary — which extends
from the southwestern part of Hardeman County, north-north-
eastward through Chester, Madison, Henderson, Carroll and
Henry Counties. As we proceed westward, the covering of
surface material becomes heavier and the clays will be found only
under deeper cover.

Fig. 8. — Digging ball clay near McKenzie, Tenn.
burden that is removed.

Shows the heavy over-

The Lagrange is not the only clay-bearing formation in Ten-
nessee and Kentucky, there being three others forming narrow
belts to the east of it, viz., the Porters Creek, Ripley, and Eutaw,
Dut these are, so far as known, of minor importance. No high
jrade clays have been developed thus far in the western part
of the Lagrange area — although ball clay is reported from Eads,
Shelby County.

458

JOURNAL OF THE

The two questions that are of practical interest arc, first ,
whether the production of these high-grade Tennessee and
Kentucky days can be increased, and second, whether any new
uses have been found for them in recent years.

Firstly, as to production. This can be increased by improved
labor and transportation facilities — a difficult matter to adjust
at the present time. It can also be stimulated by the develop-
ment of new deposits, which calls for careful prospecting. The

Pig. 9.-

-Pit of sagger clay near Henry, Tenn.
white sand.

Overlain by orange and

clays occur in lenses, many of which do not cover more than
10 or 12 acres, but some of which are considerably larger. These
lenses are invariably covered by from three to forty feet of sandy
overburden, and hence, the clay shows natural exposures only in
gullies or stream banks. The fact that another clay lens is known
to be near a developed one does not necessarily insure it being
of the same quality, indeed, even in any one pit, there may be
7 or 8 different grades of clay.

AMERICAN CERAMIC SOCIETY.

459

The indications are that a considerable quantity of undeveloped
ay may still remain in this region, but the larger proportion
better grades is probably to be found in the northern half of
ie Tennessee-Kentucky belt.

With regard to the second point — that of uses — these have
:panded considerably since the area was first developed and we
ay say that they are now used in the manufacture of white-
are bodies, electrical porcelain, glass pots, crucibles, abrasive
heels, enamels, saggers, etc.

«** -k

ig. 10. — Pit of ball and sagger clay near India, Term. This pit supplies
no less than three grades of clay and the overburden is
gravelly material.

The same general formation that carries the ball-clays of
entucky and Tennessee is found west of the Mississippi in
rkansas, but practically no development has gone on there and
affords a good field for the clay miner to prospect in. One
cception should be noted, this is near Lester, Ark., where a
id of clay averaging 5 to 6 feet in thickness is found underlying
canneloid lignite. Tests which have been made of this material

)(>< )

JOURNAL OF THE

show it to be adapted for usl: in glass-pot manufacture. There
is an excellent chance to develop more clays in this area by further
prospecting. Indeed, the formation carrying this clay extends
northward in Arkansas and southward into Texas — where the
beds of lignite are often underlain by beds of plastic refractory
clay.

In the embayment region, there still remains to be mentioned
the clays from southern Illinois which He at the northern end

Fig. ii. — Clay pit at Pryorsburg, Ky., showing no less than four grades

of clay in the face of bank. Several other grades underlie the

floor of the pit.

of the embayment. These clays have attracted considerable
attention in the last few years because of their use in glass pot
and crucible manufacture. They belong to the refractory bond-
clay type. Mining still continues but it is perhaps a little difficult
to forecast the future of the field, since the deposits appear to
lie in basins whose exact boundaries do not seem to be definitely
known.

AMERICAN CKRAMIC SOCIETY.

461

3. The Piedmont and MountainBelt. Lying west of the coastal
lain, there is a broad belt extending from western Connecticut
mthwestward to northeastern Alabama. It is a belt in which
i\v developments are taking place from time to time, but
iologically speaking, all of the clays are of the residual type,
hat is to say, they have been derived from the weathering or
reaking down of the rock to clay. These clays vary, then,
xording to the character of the original rock, which will affect
leir silica and iron contents.

Fig. 12. — Shaft for mining refractory bond clay north of Anna, 111.

At the northeastern end of the belt, we have a somewhat
olated deposit of kaolin in western Connecticut.

The next region is in southeastern Pennsylvania where kaolins
ere formerly mined, but the deposits are now exhausted although
le extension of the area in Delaware is still worked. No ex-
ansion can be looked for here. There are, however, two other
istricts in Pennsylvania deserving mention, not because they

1.62 JOURNAL OF THE

are new discoveries, but because they contain possibilities The
first of these is a belt passing through Saylorsburg, Monroe

County, and lying on the northwest side of Chestnut ridge. The
clay here underlies the Oriskany sandstone. It has been used
crude for portland cement, and in its washed form for paper.
Recent developments give promise of its being of use in pottery
and the belt should be further prospected.

It is interesting to note that white clays have been discovered
in this same formation in northern West Virginia and north-
western Virginia, but at present they are not very accessible to
transportation.

The second area is in the South Mountain district of Pennsyl-
vania, where the mountain quartzite east and southwest of Mount
Holly Springs, Cumberland County, has associated with it white
residual clays. The clays are highly siliceous and the silica is
very fine grained, so that the product needs careful washing.
The crude clay is fairly white, but it does not burn so, so far as
tests that have been made would indicate. The output, there-
fore, may be chiefly of value for paper and paints. Few com-
panies are in operation here and the mining of the clay may pre-
sent difficulties — due to the treacherous character of the ground —
but the region shows possibilities for further development.

Our next area is in Virginia, where we have white residual
clays on both sides of the Blue Ridge. Those on the eastern side
are similar to the North Carolina ones, to be referred to later,
but the Virginia deposits offer little promise — except one in the
Nelson Co. district, in the neighborhood of the titanium ore
mines.

Those on the western side form a chain of deposits extending
southward along the flanks of the Blue Ridge on the eastern side
of the Great Valley. They are usually associated with brown
iron ore deposits and may be of considerable size. The clays
which are derived from the weathering of shales, are white to
grayish white when mined, but do not burn pure white in most
cases. At present they are worked only at Cold Spring and
Buena Vista, but there are other undeveloped deposits between
these two points which are worthy of testing. The better grades
are of value for paper manufacture and possibly ceramic products.

AMERICAN CERAMIC SOCIETY

46:

Considerable attention has been given to the oeeurrenee of
kaolin in western North Carolina, northern Georgia, and north-
eastern Alabama, where deposits occur, but it is only in the first-
named state that the material seems to occur in any quantity,
and while the qualities of this material seem to be fairly well
known, the supply of future reserves is less in evidence. The
present supply could be increased with sufficient coal and labor.
Field work carried on in the region this summer shows also that

Fig. 13. — Excavation in white clay near Holly Springs, Pa. The over-
lying quartzite outcrops at the base of the slope seen in
the distance.

the field contains considerable undeveloped material, although
some of the deposits are a little remote from the railways.

With the reduction of imports, many questions are asked re-
garding the possibilities of finding kaolins in other parts of the
country. The discoveries up to the present time are few. In-
deed, in the northern states, where the soil has been combed
off down to bed rock by glacial action, there is little chance of
discovering any new ones. The deposits at West Cornwall, Con-

I'M

JOURNAL OF THE

aecticut, and St. Remi, Quebec, lying inside the glaciated area,
arc unique occurrences.

A district of white residual clay, long known but little developed,

is in southeastern Missouri around Glen Allen and Lutesvillc.
The deposits may be expected to differ in form from the North
Carolina ones because of their slightly different derivation Those
of North Carolina arc formed from pegmatite veins and hence
are long and narrow. Those of Missouri are derived from lime

Fig. 14. — Pit in white residual clay near Cold Spring, Va. This clay is
residual from Cambrian shale.

stone and hence may be extensive as to the clay mass. More
work should be done in this area, but it still remains to be proven
whether the deposits of white clay are of large extent. The bodies
are irregular in shape and rest on limestone. It is a region of
possibilities. The output at one mine is used in plaster board
manufacture.

4. The Coal Measures Region. — In our central States, and along
the mountains in some of our eastern States, coal fields of great

AMERICAN CERAMIC SOCIETY.

465

extent have been developed. Associated with these coals are
extensive deposits of shale and clay, many of them being of re-
fractory character and forming the basis of an extensive fire-
brick industry, but here and there, refractory clays suitable for
glass-pot manufacture have been developed. They are few and
far between — the most noteworthy ones being those of the
Cheltenham district around vSt. Louis. Others have been de-

FiG. 15.- — -General view of flint clay near Owensville, Mo.

veloped at scattered points in Illinois and Ohio, but one cannot
help but feel that there is a possibility of finding more.

5. The Missouri Flint Clays. — There remains to be mentioned
a peculiar type of clay found in Missouri and commonly referred
to as flint clay, but this term hardly tells the whole story.

The deposits of this material are scattered over several coun-
ties, notably Gasconade, Franklin, Osage, Warren, etc., in the
east-central part of the state. They are very numerous and
while many have been opened up, we believe that numerous others

466

JOURNAL OF THE

still remain to be found. A peculiarity is their frequent, basin-
shaped form due to the fact that the clay has been laid down
in depressions. In these depressions we therefore find this more
or less basin-shaped mass of refractory clay, often of the flint
type, with its peculiar physical characteristics. Not the least
remarkable feature is the mass of granular, oolitic clay, often
found within the flint clay, and which runs high in alumina,
even up to 75 per cent. This high alumina content is due pre-

Fig. 16. — Mass of disapore clay in flint clay deposit near Hofflin, Mo.

The rod stands in front of a mass of diaspore clay. The man in

foreground is standing on flint clay.

sumably to the mineral diaspore (A1203.H20). All of the deposits
do not carry the rough clay, as the miners call it, and the quan-
tity present does not seem to stand in any direct relation to the
size of the deposit — for the masses may be large or small. The
total tonnage visible at the present time is not large, but the total
quantity which may be obtained by the continual opening of
new pits may be considerable, and so it is a type well worth
considering.

AMERICAN CERAMIC SOCIETY. 467

Conclusions.

In conclusion, we think it is safe to predict that the outlook
for the still greater development of our domestic clays is good —
assuming that labor and transportation facilities are available.

We can also assume, we believe, that the reserves in many of
the areas referred to are also sufficient for some time to come,
and lastly, that this country can supply satisfactory refractory
clays of high bonding strength without the use of imported
materials. It has also a considerable reserve of kaolins. A single
clay may not have all the desired properties, but judicious blend-
ing of two or more clays may often yield a mixture having
the desired properties.

Cornell University,
Ithaca, N. Y.

THE EFFECT OF CERTAIN IMPURITIES IN CAUSING
MILKINESS IN OPTICAL GLASS.

By C. N. Fenner and J. B. Ferguson.

At the time when the staff of the Geophysical Laboratory began
to cooperate with the Bausch & Lomb Optical Company in a
study of the problems connected with the manufacture of optical
glass (that is, in May, 19 17) one of the matters which gave much
concern, and, for a time, was probably the chief difficulty with
which we had to contend, was the rather frequent production
of pots of milky or opalescent glass, which was entirely useless
for optical purposes.

In endeavoring to locate the source of the trouble and eliminate
it, a very puzzling feature was the apparently random manner in
which a pot of glass of this character appeared.

Among the four or five standard types of glass which were
being produced in large quantities at the time, practically the
only one which showed this phenomenon was the "light flint" —
a glass which had a PbO content of about 33.5 per cent and
whose refractive index was about 1.572.

Although the general procedure as regards the furnace treat-
ment of each pot of this type was practically the same, pots of
milky glass appeared in rather large quantities, while at the
same time the greater number were perfectly clear.

The samples or proofs which were taken at intervals while the
pot was in the furnace never showed any indication of milkiness,
and frequently the trouble remained latent or nearly so even
up to the time that the cold pot was taken out of the annealing
arch, and did not show up until the apparently clear glass was
reheated in the muffles for pressing. In fact, it might not be
noticed until the pressed glass had come out of the cooling ovens.
On the other hand in some instances, after the primary cooling
of the pot in the arch the glass appeared milky throughout, or,
more frequently, showed a slight opalescence at the top near the

AMERICAN CERAMIC SOCIETY. 469

•ircumference. In all cases in which there was any indication
)f milkiness reheating increased the turbidity, and even in those
nstances in which the opalescence appeared at first to affect but a
mall percentage of glass the whole potful was likely to become
nilky on reheating.

In endeavoring to obtain light on the matter we learned that
his trouble had appeared quite recently, apparently about the
ime that the use of Russian potash instead of German in the
>atch had been begun, although information on the latter point
vas somewhat conflicting.

The German potash, of which there had been an abundant
upply before the war and which had been used to the exclusion
)f potash from other sources, was a very pure material in every
espect. The Russian potash contained variable but always
•onsiderable percentages of K2S0i and KC1 in addition to car-
)onate. The results of some analyses by Dr. R. H. Lombard of
he Geophysical Laboratory are shown in Table I.

Table I. — Analyses of Russian Potash.

SO3. Cl. H2O. K2CO3.

0.75 per cent

1.9 per cent 7 .02 per cent

; 3.6 per cent

2 .47 per cent 2 .03 per cent

1 12.2 per cent 3 .4 per cent 10 . 1 per cent 56 . 1 per cent

» 7.7 per cent 3 .6 per cent 7 .5 per cent 68 .7 per cent

In making up the batch about 68 parts by weight of potash
calculated to anhydrous K2CO3) to 300 parts of sand were used.

While the cause and prevention of the milkiness were still
inder discussion, Dr. H. E. Merwin drew our attention to an
irticle which appeared in the Journal of the Society of Glass
Technology,1 in which very analogous phenomena were described.
I neir appearance was attributed to the presence of chlorides and
sulphates (especially the latter) in Russian potash, and experi-
nents in confirmation were carried out. Evidence was obtained

1 J. D. Cauwood and W. E. S. Turner, "The Influence of Small Quanti-
ses of Chlorides and Sulphate in Producing Opalescence in Glass," J.
Soc. Glass Tech., 1, 187 (May, 1917).

470 J< lURNAL OF THE

by Cauwood and Turner that a high temperature and extended
period of ti nc in heating had beneficial results.
It seemed reasonable to suppose that our phenomena wen- to

be accounted for in the sane way, but certain considerations
rather obscured matters, hirst: there did not seem to be any
very direct connection between the amount of sulphate or chloride
in the potash and the tendency of the glass to become milky.
Second: a number of runs were made in an experimental furnace
under conditions closely similar to those in the large furnaces.
Groups of six small pots holding a few pounds of glass each were
placed in the furnace simultaneously. Various amounts of sul-
phate and chloride were added to the batches. In a given group
some pots showed milky glass and others did not, and here again
there seemed to be no relation between milkiness and amount
of sulphate or chloride added. Third: during a laboratory study
of the ternary system CaO-MgO-SiO? by one of the present
writers (J. B. F.) it had been found that with glasses having
compositions within the SiC>2 field, especially with those not so
high in Si02 as to bring the composition far from the two-phase
boundary, a similar milkiness had often been obtained when the
crucible was lifted from the furnace (above melting temperature)
and quenched in water, especially if the transfer were not carried
out very rapidly. At times the depth of milkiness was so great
that the glass looked like the so-called Carrara glass. In such
instances it had been possible to detect with the microscope
multitudes of minute crystals of Si02. This phenomenon could
not be ascribed to the presence of chloride or sulphate — as the
"ingredients employed were" of a very high degree of purity.
Therefore, it was evident that milky glass could arise from other
causes.

The opinion toward which we leaned, in view of the evidence,
was that sulphate and chloride were probably responsible in some
way, but the remedy was not plain. It was not considered
practicable to run the furnace temperature much above 14000 C,
at which point it was supposed to be held at the time, without
running much risk of corrosion and perhaps failure of the pots,
nor was it possible to add carbonaceous material to these lead
glasses to transform sulphates and chlorides to carbonates.

AMERICAN CERAMIC SOCIETY. 471

evertheless, before the cause of milkiness was definitely de-
rmined, a remedy was found, and later there was a gradual
cumulation of evidence as to the cause.

In a previous paper1 the senior writer has described the work
lich was done at the Bausch & Lo.mb plant preliminary to in-
dling a means of temperature-control of a more reliable character
an the thermoelements which had formerly been in use. This
>rk showed that the readings given by the thermoelements
.ried widely from the true reading and, moreover, that the
(Terence was not constant. Therefore, it became evident that
e temperature treatment of different pots had been by no
eans as uniform as supposed. A Leeds and Northrup optical
■rometer was substituted and personal attention was given for
yeral months to keeping the temperatures correct. The furnace-
m were instructed in the use of the instrument, so that tem-
ratures at night might be controlled by it, and in most cases
pendence could be placed on instructions having been carried
t. The result was that from that time forward for a long
riod there was scarcely a pot of milky glass produced, and when
:er a few appeared the reason was fairly obvious.
We took special pains to see that the temperature during
dting and fining, that is, for a period of about 36 hours, was
Id at 14000 C or a little higher. The use of Russian potash
is abandoned and in its place a much purer material supplied
' the Armour Fertilizer Works was used. For the latter ma-
rial specifications required that the SO3 content of the potash
ould not exceed 0.3 per cent and that the CI should not be in
cess of 2.0 per cent. Analyses were made of the contents of
ch barre supplied and material exceeding these limits was
bject to rejection. In general, the furnace temperatures were
Id at about 14100 C, in order to be on the safe side, and if the
)3 or CI content was near the limits considered safe, the tem-
rature was run up to 14200 C. With this increased care and
rtainty of control, not only was the danger of milky glass al-

1 C. N. Fenner, "Methods of Temperature-Control in Glass-Melting
irnaces," Phys. Rev., 2nd series, II, 2, 141. Abstract of a paper presented
the Rochester meeting of the American Physical .Society, Oct., 19 17.

i, • JOURNAL OF THE

most eliminated but also the loss of pots through leakage was
very materially lessened.

For some time after the installation of this system, frequent
tests were made of the tendency of the glass to become milky 1>\
heating small pieces for several hours in clay crucibles to a tun
perature of 950-10500 C in an electrical resistance furnace.
It was found that in many instances glass which passed through
all ordinary manufacturing processes without any indication of
the appearance of milkiness still held sufficient latent possibilities
in that direction so that the experimental treatment mentioned
developed it to a point where it became visible. Study of the
results showed in many cases that this development of opalescence
or milkiness could be correlated with the fact that the potash
used in the batch contained relatively large amounts of sulphate
or that for some reason the furnace temperatures had been a lowed
to drop below the established point for a considerable length of
time.

At a period considerably later than that to which we have been
referring it seemed desirable to cut down, if possible, the length
of time for which pots remained in the furnace. Apparently
as a direct consequence of this change pots of milky glass began
to appear again. Whereas under the old schedules a potash
containing 0.30 per cent SO3 could be used without danger and one
containing even 0.70 per cent SO3 could be used without too much
risk if the temperature were raised to 14200 or 14300 C, under the
new schedule (about nine hours shorter) milkiness appeared when
potash containing 0.30 per cent SO3 was used in the batch. The
trouble was much lessened, however, when the length of time o
stirring was increased.

All the evidence leads to the inference that the trouble is to be
ascribed to the presence of very small amounts of SO3 and, to a
less degree, CI in the melt. Apparently these substances, when
present, are continuously eliminated, but the evolution of the last
traces is a very slow process. The elimination is favored by a
high temperature and prolonged time, as Cauwood and Turner
found, and also by the agitation of the melt which accompanies
stirring.

"

AMERICAN CERAMIC SOCIETY. 47.3

If glass which is milky or is likely to become milky is heated
noo-11500 C the milkiness disappears, but glass so treated
velops numerous bubbles, which render it unfit for use, and a
;ser degree of heating does not decrease the milkiness.
Several analyses (made by J. B. F.) show the actual S03-con-
nt of the glasses. In one case a very milky glass contained
14 per cent S03. After being cleared up by heat-treatment at

00 ° C or a little higher, and having thereby lost its tendency
j become milky, the same glass contained 0.06 per cent S03.
In another case two lots of glass were made at the same time
the same furnace (a two-pot furnace) and from the same batch.

robably as a result of unequal distribution of temperature in
le furnace, one pot became milky and the other did not. The
ass from the first pot was found to contain:

S03 0.10 per cent

CI 0.17 per cent

The glass from the second contained:

SO3 o . 05 per cent

CI 0.17 per cent

robably n general the presence of CI in the potash is less serious
lan the presence of SO3 when the heat-treatment is similar to
tat which we followed, but at times enough CI may be left to
rengthen the effect of SO3 and cause milkiness to appear in a
ass which would otherwise remain clear. The last case may be

1 example of this.

Although the origin of milkiness in glass in the cases described
:ems to be connected fairly definitely with the presence of SO3 or
1, it is doubtful whether the minute particles to which the milki-
jss is due are themselves composed of sulphates or chlorides.
irect evidence on this point is very difficult to obtain. With
le best microscopic magnification it is barely possible to discern
iserete particles of foreign matter and nothing can be made out
igarding their properties. In fact, the bluish color of the
Dalescence is itself evidence that the particles are so minute as
) be of a similar order of magnitude to the wave-length of light
id thus produce a scattering of light rays. According to Wood1
1 R. W. Wood, "Physical Optics," 1911, p. 624.

.174 JOURNAL OF THE

11 a beam of light 's passed through a transparent mediui
containing in suspension small particles, the refractive index o
which dilTers from that of the surrounding medium, light will bi
given off by the particles in all directions. In the case <>i particle
of the order of magnitude of the light waves, the amount of ligh
scattered increases as the wave-length is decreased, which explain
the preponderance of blue always observed in these cases.'
However, it was found by Mr. II. S. Roberts, of the Geophysica
Laboratory, that the size of the particles may be increased b]
holding a piece of glass of this character at a softening tempera
ture (say 900 iooo0 C) for several days. Dr. H. E. Mcrwir
examined microscopically the glass so treated and was able t<
determine that the particles now showed an angular form an<
possessed a refractive index considerably below that of the glas
in which they were imbedded. It seems most probable that th<
action of the sulphate and chloride impurities is of the sort t<
which the term "catalytic action" is applied for lack of bette
knowledge. Ordinary glasses are undoubtedly in a state of un
stable chemical equilibrium at the temperature at which the fina
stages, at least, of the furnace work is carried on. Their stabli
condition would be one of partial devitrification, and it is ai
astonishing fact that the unstable state is retained for a pro
longed period under conditions of treatment which seem favor
able to crystallization. Whatever may be the factors whicl
cause the glassy condition to be retained it appears that th
presence of sulphates and chlorides, in quantities so minute tha
the physical properties should apparently be in no way affected
nevertheless does cause such a change as to permit swarms c
sub-microscopic crystals to develop. The phenomena exhibite
are of a kind for which very little real explanation can be offeree
In geological processes it has long been recognized that th
presence of certain volatile constituents, termed mineralizers,
especially favorable to the segregation and crystalline develot
ment of minerals in rocks, and the matters here described are a]
parently of a closely similar nature.

Another occurrence at the Bausch & Lomb plant is of intere
in this connection. The light crown glass which was made at tl
works was never known to be affected with milkiness and it w;

AMERICAN CERAMIC SOCIETY 475

pposed that the Russian potash on hand could be employed
ithout detriment in making glass of this type. In a case in
hich this was attempted the potash contained 6.5 per cent
33, 3.6 per cent CI and 7.5 per cent H20. During melting the
lpurities separated out and formed a layer of "salt water" on
le surface of the glass. At a later stage, during cooling, the glass
scame so viscous that stirring had to be stopped and the pot
ithdrawn. When cold the resultant glass was found to con-
in quantities of nodules or spherulites of devitrified material,
hose general diameter was four or five millimeters, although some
asses of much larger size were present. Microscopic examina-
on of this material shows that it has a refractive index of about
485, very low birefringence, and the typical branching skeleton
rms of cristobalite — the high temperature form of silica.1 A
^termination by Dr. R. H. Lombard gave a content of 0.28
it cent SO3 in the glass, and the conditions seemed to be favor-
)le in this instance for the glass to retain the maximum amount
' SO3 possible. The small amount present acted as a very
ficient aid to devitrification.

An analysis made by Dr. E. T. Allen on a glass for searchlight
lirrors manufactured in England, gave 0.74 per cent SO3 and
10 per cent CI. The other important constituents were

Per cent

Si02 71.34

CaO 14.73

Na20 10.37

K,>() 0.44

his glass showed no indication of milkiness or devitrification.
khis fact and other evidence given show that the devitrifying
feet of sulphate and chloride varies with the composition of the
lass.

Although in the cases which have been described the evidence
idicates that S03 and CI were responsible for the production of
lilky glass, it is quite certain that these are not the only sub-
:ances which are likely to give rise to it. The peculiar "mineraliz-
lg" role played by SO3 and CI is probably filled in some cases
' C. N. Fenner, "Stability Relation- of the Silica Minerals," Am. J.
"■, (4] 36, 355 (Oct., [913).

476 AMKRICAN CKRAMIC SOCIETY.

by AS2O3 or CaF2, and there may be other substances having
like effect. Also some types of glass (probably quite different
in chemical composition from ordinary commercial glasses), such
as the CaO-MgO-SiOo glasses mentioned on an earlier page,
have a tendency toward devitrification in this opalescent or
cloudy form, which is apparently inherent in them and does not
arise from impurities.

Summary.

In the manufacture of optical glass at the Bausch and Lomb
plant, a matter which gave considerable difficulty for a while
was the occasional production of pots of glass which were affected
by opalescence or milkiness. The evidence indicated that the
source of the. trouble lay in the sulphate and chloride content
of the Russian potash to which recourse was had when the German
supply was cut off, although certain facts tended to cast doubt
upon this conclusion. It was found, however, that the trouble
disappeared when more reliable methods of temperature control
were installed, by which an assurance could be had of keeping the
temperatures constantly at 1 400-1 420 ° C, and when the Russian
potash was replaced by an American product, more nearly free of im-
purities. Later, evidence was obtained which connected the
milkiness quite definitely with the impurities mentioned, at least
as regards the case under discussion, although in other cases the
same effect is to be ascribed to other causes.

Reasons are given for the conclusion that the milkiness is caused
not by the separation of sulphates or chlorides themselves but t(
some slight change in the physical properties of the melt whicl
permits the separation of clouds of minute crystals of cristobalite

Geophysical Laboratory,

Carnegie Institution op Washington,

Washington, D. C,

September 25, 1918.

SILICA REFRACTORIES.1

By Donald W. Ross, Pittsburgh, Pa.

Previous Work.

\ great deal of work has been done on the stabilit} relations
the silica minerals, the results of which are probably sum-
med best by Fenner.2 More recently, the results of these studies
se been applied in the optical examination of silica brick.
^Dowell3 has been one of the foremost investigators along
s line, and at the present time Insley4 is conducting an ex-
lstive study of silica brick made from practically all of the
ding varieties of materials used commercially in the United
ttes for this purpose.

Vhe present status of our knowledge along these lines is briefly
follows: The specific gravity of quartz is 2.65, that of cristo-
ite is 2.33, and that of tridymite, 2.27. Upon heating, alpha-
irtz (the stable form at atmospheric temperatures) is inverted
beta-quartz at 575 ° C. In the presence of a flux, this beta-
irtz is transformed into beta-tridymite at 8700 C. This, in
n, is transformed at 14700 C into beta-cristobalite — which
the stable form from this temperature to its melting point,
50 C. "In the absence of a flux, the beta-quartz is trans-
tned directly to cristobalite — which in this case is the final
m."5 In the manufacture of silica brick two per cent of lime
lO) is mixed with the quartzite. Upon heating, the lime

1 By permission of the Director, Bureau of Standards.

2 C. N. Fenner, "The Stability Relations of the Silica Minerals," Am.
>'«'., 36, 331-384 (1913). In this connection it is of interest to note that
k recently published by the Geophysical Laboratory of the Carnegie
;itution, indicates that the melting point of pure cristobalite is probably
50± io°C instead of 1625 ° C as previously reported.

3 J. Spotts McDowell, "A Study of the Silica Refractories," American
:itute of Mining Engineers, Bvll. no, p. 1916.

4 H. Insley, U. S. Bureau of Standards, Pittsburgh, Pa.
6 Am. J. Set., 36, 339 (1913).

)7s JOURNAL OF THE

combines with a small amount of the silica to form a glass, 01
in some cases, possibly a little mono calcium silicate (CaO.SiOfl
The quartz of A silica brick is transformed first to cristobalitc
but as the lime constitutes a small amount of flux, continues
heating causes the silica to be slowly converted into tridymite-
after approximately two-thirds of it has been transformed inti
cristobalite.

McDowell1, Bell,2 Le Chatelier8 and others have also stndie(
the difTerences in finished silica wares due to the use of the differ
ent quartzites and to variations in the methods of manufacture.

Usual Method of Manufacture.

Ordinarily, silica brick are manufactured by crushing thi
quartzite so that it will pass through a 2 -inch ring. The crushe(
material is then fed in batches into wet pans and ground unti
the largest particles are approximately 0.2 inch in diameter
during which period two per cent by weight of CaO as milk 0
lime is added along with enough water so that the resulting mi:
can be molded into bricks. When dry, the bricks are burned ii
fire-brick kilns of the usual type.

Nature of the Present Work.

Some three years ago, the writer began a study of the silic
refractories under the direction of Mr. A. V. Bleininger. Th
present paper is an attempt to set forth a summary of the dat
obtained since that time and to present some applications of th
same. In taking up the data- we shall first consider the raw ms
terials, then manufacture and burning, and lastly some of th
properties of the burned ware.

Raw Materials.

By far the largest part of the silica brick in all countries ai
manufactured from quartzites of comparatively early geolog
age, although in some cases, chalcedony, chert, quartz sand, 1,
pebbles cemented together with chalcedony, have also bet

1 hoc. cit.

- C. K. Nesbitt and M. F,. Bell, "Silica Refractories," Proc. A
Sac. Testing Materials, 1917.
3 Rev. metal., June, 191 7.

AMERICAN CERAMIC SOCIETY. 479

:d. Suitable quartzites range from the hard, highly nieta-
irphosed varieties having tightly interlocking grains, to the
dium soft rocks which are slightly porous. Loose grained
artzites and sand stones are, however, of little value. The
artzites now being used range from the very young rocks to
j oldest sedimentaries.

fhe following microscopic method may prove of value in judg-
; the suitability of quartzites for the manufacture of silica
ck. In the metamorphosis of sandstones to quartzite, the
*ks which received but little alteration consist of hard,
lally rounded grains of quartz imbedded in a ground mass of a
ter material. If a thin section of a rock of this kind is viewed
der the microscope, the softer material appears as cloudy
:as between the quartz grains. These cloudy areas consist
fragments of silica (probably quartz) and impurities such as
drated iron oxide. As the metamorphosis proceeded, the in-
stitial silica, frequently augmented by silica from water solu-
n, crystallized on the larger grains and shows the same orienta-
n as these grains. Silica crystallizing in this way is pure—
;ept for occasional inclusions. Hence, the impurities are
negated. In thin sections of highly metamorphosed quartzites
;se segregations appear as sharply defined lines between the
htly interlocking, built-up grains of quartz. The quartzites
ich have been found most satisfactory for the manufacture of
ca refractories are those which have moderately to tightly
erlocking grains, while those materials which merely show
udy areas between the harder grains have been found to be
s desirable. The specific gravity of quartz is 2.65. The
icific gravity of chert rock — although practically the same as
it of quartz — is usually slightly below this figure as shown by
). 29 (Table 1) (2.585). This is borne out by Mellon1
Screen analyses (by the wet method) and porosity determina-
ns were made on 15 leading varieties of unburned silica brick,
lese are presented in Table 2, together with the porosities and
^cific gravities of bricks of the same varieties which have re-
ved regular commercial burning.

1 J. W. Mellor, Trans. Eng. Ceram. Soc, 15.

48o

JOURNAL OV TlIIv

«j

a

>>

TJ

■c

c

o

H

-

JD .

a

0 CI

a

H

v- a

0

o

■j

o

o

CX c

a

£ "°

SI

=«

IN

r^

>o

3

■*

-t

i*J

iO fN r*5
oo r>. r-.
iO (N fN

CN fN (N

o

10

a*

vO

*C

fN

(N

B **

0 0
>0 o

.* o o
Ogg

B o o

0<«O

•- CI

£

+•9

.J

.M

!5

u

■m a

o

O

W

a H

-

CI

»*ri

15

0

.O

B IN S3

B

3

E

0
u

5 £

cr

O-o

-<N JS

B
CI

B
F

B 0

.* u

H

"8 "

E w

u .2

B

E

B

c.

U

H

<

Ch

B >

o-o

oE

K o

S B

CI C

S c

<5>-<

3

E c

0 o

AMERICAN CERAMIC SOCIETY.

481

>o

L"

^•10 —

— ■ — CM —

00 *0 (N
CN (N

-t

gs

CMOCOOCMf^CMOOO^OOrO

000 O^

CO 0^ NO — OvO\OfNffJ(N^O —

r^ O <N

<N

t£

^asoooooM^.CMpr:*oi""-<Noo

O
(N rf rr

— Or^oor^io^fOfNfN — (N

(N (N

X

V-

>o r^ \0 r>- On (N <n 0* - r^iooo

10 O rr

mrN.iT>tOTt-fNtNr0r/)(N00t>
CM — —

<N (N

■<*•

BS

^ — IT

CM — — —

u-j r^ <N

6§

O

<no ■*

(N 00 r*-

OMOOr^l^CMCMrOCM — CM ■<$•

— <N(N
<N<N

m

6§

CMccf-^ooro^oo^Ov^ooc":

ONM
<N u"> T*-

— ■^OOMOvOrOf'iCMCM'Ocr:

rsrs

r^

6§

rfCM — — vO — On^ — OCMO

-OOCM^N'tfO't^O-

lOOO <N
(N CM

(N

65

— 00 PM

rMTfCN^OONfjiorf'CNCMMO —

vO "0 fN
(N(N

vO

68

vO 00 N ro O ^O "". — (Nr^Tt-CO

r-. r^ P'

— ON — f^OCO- 'fJi'fOO'.M

vO vo r-j

u

a

c
u

E
0

Porosity, raw brick
Porosity burned brick
Sp. gr , burned brick

6
v.

■

.s:

n S

h

5 0

v.

^oooocoboooo
— — cm rsi

bo'C
j: °

4) -Q > V

rt O uh

5*1

00»«)--00000

— coooocooo

0000000000

rt 3 ■

O C u
<u c o

6S

O 4/

0! <
M Z

•^iocmvCCmO^-coCMOnnO^O

n00'OONO>O00r000^''O

es

O — fN^O^O— • CO co r*) <N 0\ CO

&8

rO'tO-NT;n^rN^'t\c
0"")|OCM'<tr*»CMCNCMCM- —

rfOMOOON cO C> nO ^0 r» r-

loO'-OvGoooi-o-O'on

i-

(N-'0>0000-'t(NOlO

f-or^ioioON^fNoorsi^PNio

0^*^f*5^*fNf^rOf^lO — CM

BS

<toooo 0000 000

■ s:

(NfOi'^OOO'OOO

~ -h CN CN

u c
t/3 0

2 «

z

JS °

1 Hto

&££

5 O

S 2

r- • S"

. a m o

2 st;

o u B-s »

p§°aS'g

■2.2&.S&
u- » u I

o

is. JOURNAL OF THE

Curves showing the results of screen analyses made on quartzites

that have been crushed ready to be molded into brick, when
compared with the theoretical curve for minimum pore space,1
show that there is always an excess of material of the sizes which
correspond to those of the quartz grains.

In crushing, sandstones and quartzites that are but slightly
metamorphosed break mostly to individual sand grains and
yield an excess of these sizes.

Moderately metamorphosed material, i. c, material that is
still slightly porous, follows the theoretical curve quite closely. In
this case the breaking is usually through the interstitial material
between the built-up grains. However, the interstitial material
breaks free from some of the original rounded grains and still
other grains are broken in two.

In the case of highly metamorphosed quartzites, i. e., those in
which the crystalline quartz is practically continuous, the break-
ing is usually through the grains, but a fair percentage of rounded
grains also break free from the interstitial material. This break-
ing away of the interstitial material from the rounded grains
is particularly noticeable in the highly metamorphosed Baraboo
quartzites.

Inspection of the screened material revealed the fact that in
the Medina, Baraboo, Homewood, Missouri, and Eastern Penn-
sylvania materials, the original rounded grains which broke free
from the interstitial material practically all passed through a
30-mesh (0.503 mm. diam.) screen, and that but very few parti-
cles passed through a 60-mesh (0.221 mm. diam.) screen — th<
great majority being caught on the 40-mesh (0.381 mm. diam., I
and 60-mesh screens, in varying proportions. In Fig. 1 an
shown photographs of samples of material remaining on the 40
mesh screen. No. 6 consists largely of material broken througl
the interstitial material between the built-up grains, and sorm
that has broken through the grains. A slight amount of finel;
ground material still adheres to the particles. No. 21 shows
large percentage of the original rounded grains which have broke
free from the material of the built-up grains. No. 12 contain
a large percentage of material broken through the grains an
1 Taylor and Thompson, "Concrete, Plain and Reinforced," p. 775.

AMERICAN CERAMIC SOCIETY. 483

I
so material in which the breaking was between the original

mnded grains and built-up grains. A slight amount of finely

round material still adheres to the particles. In No. 14, the

riginal grains of the material appear more angular than is the

ise with the other quartzites examined. In No. 31, many

Fig. i. — Materials (caught on 40-mesh sieve) resulting from wet screen

analyses of raw commercial silica brick mixes. Magnified.

(Microphotographs by Dr. R. Thiessen, U. S. Bureau of Mines,

Pittsburgh, Pa.)

f the original rounded grains have broken free from the chal-
idony bonding material. In No. 29, the grains in the chert
re extremely fine, hence, as might be expected, it breaks into
tigular fragments. A slight amount of finely ground material

IM JOURNAL OF THE

still adheres to the particles. The three grains cheeked are of
a different material which was accidentally included.

At the present time, McDowell is carrying on a study of the
si/.es of the built-up interlocking grains of several quartzites.

Of the quartzites, the loosely bonded varieties yield the most
porous raw brick, and the strongly bonded varieties, in which
the breaking is across the grains — thus avoiding an excess of
material corresponding to the grain sizes — yield raw brick of the
least porosity. Material No. 12 (Table 2) is of this type and has
the lowest porosity of any of the unburned brick tested. The
chert material, No. 29, is next, this being closely followed by the
mixes made from Baraboo quartzite, while the large majority
of the raw mixes made from the Medina quartzite, of Blair and
Huntingdon counties, Pennsylvania, show in turn somewhat
greater porosities than do the Baraboo materials.

Innovations in Methods of Manufacture.

Finished silica brick frequently show quite wide variations in
their physical properties when subjected to strength tests, etc.
These variations may be due in part to differences in grind from
one pan to the next, and, in part, to the human factor in hand
molding. To overcome such difficulties in special cases, the fol-
lowing changes in the methods of manufacture have been sug-
gested :

The material is ground and, by screening, is divided into three
sizes. These are then mixed in the desired proportions in a
machine similar to that used for mixing concrete. To avoid
storage in bins,1 the material might be ground directly in a wet
pan having raised mullers. The lime is added at this point.
The brick are then machine molded. This should insure greater
uniformity of texture, shape, amount of material per brick, and
should tend to slightly reduce the porosity. Nesbitt and Bell2

1 When crushed materials are poured from a spout into a bin there is a ;
tendency for the coarse material to segregate from the fines. This trouble
may be largely obviated by frequently changing the position of the spout
while a bin is being filled. It is essential to keep this in mind when using
bins for the storage of ground silica brick materials.

2 Loc. cit.

AMERICAN CERAMIC SOCIETY. 485

have suggested that 1500 pounds per square inch is the most de-
sirable pressure for use in the manufacture of machine molded
silica brick. Fine surface cracks appear to be characteristic of
machine-pressed silica brick. However, in most cases these should
be of no serious detriment to the bricks. In making silica brick
by machine, the water content need not be quite so great as for
the hand-made brick. However, an abnormally low water
content is apt to result in an open brick of weak structure — both
in the raw and burned condition.

Effects Produced by Burning.

Since the specific gravities of cristobalite (2 . 33) and tridymite
(2.27) are considerably lower than that of quartz (2.65), the
relative position of the specific gravity of any sample of silica
brick between these limits should indicate, quite accurately,
the degree to which its quartz has been transformed to the lower
specific-gravity forms.

To accurately determine the specific gravities of solids by
means of a pycnometer bottle, using finely powdered material,
is a rather tedious operation. Hence, an effort was made to
determine the specific gravity by using a two-inch cube of the
material. This was first weighed dry, then saturated with water
and weighed wet, and finally the saturated piece was weighed,
suspended in water. From these data, the specific gravity was
calculated. Saturation was obtained by placing the pieces in
boiling water and subjecting them, while thus submerged, to a
vacuum equivalent to twenty-four inches of mercury for a period
of four hours. This method, when carefully conducted, has
been found to give the specific gravities with sufficient accuracy
for our present purpose.

The specific gravities of most of the leading varieties of com-
mercial silica brick, manufactured in the United States, have
been determined in this way (see Table 3) and, thus far, all lie
between 2.65 (the specific gravity of quartz) and 2.27 (the
specific gravity of tridymite). The results of our own work,
and more recently of that of other investigators of silica refrac-
tories, indicate that this gives a definite and satisfactory indica-
tion of the degree to which a brick has been burned.

486

!< >URNAL OF THE

Table 3. -Porosities and Specific Gravities of Principal Brands op
Commercial Silica Brick Mads in the United States.

QuartzitC from which the lirick wire made

Sampli

No.

Specific
gravity.

I'c 1 ie ill
porosily.

I

2-375

27-15

2

2 .296

2584

3

2.424

28.52

4

2 393

2 7 . I 6

5

2.340

22 .80

6

2 .291

26.35

Medina (Tuscarora) \

7

2-375

28.35

8

2 336

31 52

9

2.321

25 97

10

2.358

26 . 10

1 1

2 .291

29.56

12

2 .460

2310

I '3

2.340

29.88

Aw, 2 .357

Av., 27 .07

M

15
16

2.496
2.448
2517

30.58
28.95
26.55

Homewood sandstone.

Av., 2.470 Av., 28.69

Oneida.

17

2.274

25.66

Baraboo.

18

2.381

23.91

19

2.43O

25-I5

20*

2.325

31.41

21

2 .390

23-45

22

2-395

24.79

Av., 2.399 Av., 24.30

23

2-495

23 .22

Quadrant formation (Montana) i

24

2.490

22 .66

1 25

2-537

22 96

Alabama, probably Weisner formation
Dakota and Comancl

Indiana chert

Saint Louis district..
Eastern Pennsylvania

Grand average of all varieties
* Calcined quartzite.

26

2 .311

29-75

Dakota and Comanchian (Colorado) . .

/ 27
I28

2-393

2.387

24.63
2451

Indiana chert

29

2.273

25.64

Saint Louis district

f 30

131

2 363
2393

25-73
24.70

32

2.335

26.85

Av., 2 .384 Av., 26.34

AMERICAN CERAMIC SOCIETY.

487

Using the specific gravities of the materials in this way, our
data (Table 4 and Fig. 2) indicate that at the end of heating
to 15000 C (conducted as follows: heating to 8oo° C in 18 hours,
heating from 8oo° C to 15000 C in 6 hours, and held at 15000 C

■

A/umber

t

of TTmes Burned

} 4-

s

6

7

4C

2H5.

— -££.

2 60

/

V

V

Gn

- Specific C

u

/

*

Group2 = Percent fxt,Vo/f*p./r>7>
6roup 3 -Per cent Solid Vol lip, tern

isCr/yJtffot^
UOriaJrtVoi

sv;

A

pso.

P= Pennsylvania^ edino Quart zif

fc>

to

A-

■ Alabama, Weisner Quartz/

te

240

2/

A>£

-B
^P

3

230

0

1300 /350

Tern perafure in

1450

Fig. 2. — Changes in exterior and solid volumes and specific gravities of

Medina, Baraboo and Alabama quartzites, caused by repeated

burnings to successively higher temperatures from

1200 "-1500° C.

for 1 l/2 hours) the amount of transformation that had taken
place was approximately the same in the Baraboo, Alabama,
and Medina quartzites. The specific-gravity curves show
that, for the repeated heatings at from 13000 to 14000 C,
the Medina quartzite appears to be transformed slightly
more than the other two. The average specific gravity (Table 3)
of commercial brick made from the Baraboo quartzite
is somewhat higher than that of brick made from the
Medina quartzite. Hence, since the chief part of the burning
received by average commercial brick is between 13000 and 14000
C, it is probable that at these temperatures Baraboo quartzite
is actually transformed at a slightly slower rate than is the Medina

|S.s

JOURNAL OF THE

Table i Specific GnAvmsa o* Quartzitbs bbporh and aptkk Heating

tor i l/i Hoiks at 1500° C.

Specific Kr:>vity Specific univity "f quartz-
Material, Raw quartzite, fte aftei 1500 C burn.

Medina quartzite 2 .635 2 .303

Baraboo quartzite 2 .650 2 .293

Alabama quartzite 2 .640 2 .295

Homcwood sandstone 2 .630 2 .326

Montana quartzite 2 .638 2 .306

quartzite. When quartz is transformed to eristobalite, the action
is progressively from the surface and along cracks towards the
interior and hence is most rapid in the varieties having the great-
est surface areas. In accordance with this we would expect the
slightly porous Medina quartzite to be transformed somewhat
more rapidly than the more highly metamorphosed, non-porous,
Baraboo quartzite. At temperatures above 14000 C, however,
the transformation of quartz to the lower specific-gravity forms
is so rapid in all varieties that the above effect is apparently ob-
scured.

In silica brick, the transformation from eristobalite to tridy-
mite apparently takes place first in the fine material of the ground
mass and then progressively from the surfaces to the centers of
the larger particles. This is illustrated in Fig. 3, which shows a

Fig. 3.— Brick made from Medina quartzite, after 40 heatings to 14500 C.
Ground mass is tridymite, gray outer portion of large particles is
tridymite in glass, and white centers of large particles are
eristobalite (natural size).

AMERICAN CERAMIC SOCIETY. 489

fractured surface of a brick which has been burned to an advanced
stage. In this case the quartz has almost entirely disappeared
md practically all of the cristobalite that remains is represented
3y the light patches present at the centers of the larger parti-
cles. In the course of the transformation of quartz through
cristobalite to tridymite, the original crystal form is largely lost,
so that the final product shows a considerable quantity of inter-
locking crystals which were not present in the unburned brick.

In accordance with Mellor,1 we found that chert is trans-
formed to the lower specific-gravity forms much more rapidly
than quartz. Thus, brick 29 (Table 3) (micro-analyses shown in
Table 1), which was made from a chert rock, received one regular
burn in a commercial silica brick kiln. Its specific gravity is
lower and the sum of its contents of cristobalite and tridymite
is much higher than is usually the case with single burn bricks
made from quartzite.

A comparison of the porosities of raw and burned brick (Table 2)
shows that the porosity of the burned brick depends largely
upon the porosity of the raw mix. However, it was found, by
repeated trials, that if a raw brick mix was heated to 15000
C in one day the resultant product was porous, weak and friable,
while if fourteen days were consumed in reaching the same maxi-
mum temperature, with a protracted soaking at from 12000 to
13500 C, the resultant brick showed no undue increase in porosity
and was sound and strong. In fact, in cases where the conversion
of quartz has been practically completed at the soaking tempera-
tures, the final increasing of the temperature to the maximum
actually causes the porosity of the brick to decrease. As a rule,
the increased porosity (punkiness) of rapidly burned brick is
much more pronounced with loosely than with firmly bonded
quartzites.

Number 37 (Table 1) represents a piece of brick which received
long soaking in use in a by-product coke oven at moderately
high temperatures. The differences in micro-composition from
the hot (heating flue) side of the brick to the cold (coking chamber)
side are very marked. For the specific gravity and porosity
determinations, pieces were taken which extended entirely across
1 Loc. cit.

490

JOURNAL OF THE

Slum bet of Imic Burned

260

/__

M

*

/ - Specific Oray/fcf

y

^250

2Z E*. fenor Volume En pan s
3 - Solid Volume Ex pan si

on

on

?n

">>s/

240

\ 2^

'0

^f^

%**

230*

r^ '*

0(

1100

1750

1300 1350 1400 1450 /SOO

Temperature in "C
Fig. 4. — Changes in exterior and solid volume and specific gravity of
Medina quartzite, caused by repeated burnings to successively
higher temperatures from i2oo0-i5O0° C.

1

Number of Times Burned
2 3-4

s

0

r

1- Specific Groritu
2- Exterior Volume Expansion
3 - Solid Volum e Exp an si on
4- Per cent Porositu

\l

^

.* *

s

^2 JO

^^^"

Si

%

%

/

240

r ^^^

2^~

2.m

^s£^-^—T

30 h

I20O 1250 1300 1350 /400 I4S0 1500

Temperature in "C

Fig. 5. — Changes in exterior and solid volume, specific gravity and

porosity of Medina raw mix — caused by repeated burnings to

successively higher temperatures from i20o°-i50o° C.

AMERICAN CERAMIC SOCIETY. 491

ion of this brick. The abnormally low porosity of this
was in all probability caused by the protracted soaking at
emperatures which it received in use. This brick is of
liar interest because of its low cristobalite content, even
:oke side— where there is considerable quartz + silicates
t.

omparison of the specific-gravity curves in Figs. 4 and 5
tes the way in which the lime increases the amount of
>rmation produced by a given heat treatment. Increas-
e lime content above two per cent appears to further in-
the rate of transformation. Sodium chloride apparently
similar effect. On the other hand, plastic clay, pure
:d alumina (ground to pass through a 60-mesh screen),
on oxide appear to have very little effect on the rate of
Donation. Small percentages of fluorspar (CaF2) were
n conjunction with lime, and although our data is not con-
I the indications are that the fluorspar has very little effect
5 rate of transformation.

Properties of the Burned Brick.

results of cold cross breaking tests (Table 5),1 in con-
3n with the natures of the materials broken, and their
c gravities, porosities (Table 3), micro-analyses, etc., indi-
mt, with brick made from quartzites, the strength of slightly
derately burned brick probably depends largely upon the

Table 5. — Cross Breaking Tests. (Modulus of Rupture.)

Material.

(18) Baraboo. (6) Medina. W. Va. (29) Chert. Medina.2

565

1557

664

593

709

"43

544

452

782

546

1003

645

527

863

809

446

Av 555 91.5 809 ... 793

Jricks set on edge with six inches between supports.
Vs given by McDowell.

|.,' JOURNAL OF THE

state of metamorphosis of the raw material and the- porosity of
raw brick and the lime-silica bond formed in the burning,
the other hand, well burned silica brick appear to have 1
strength somewhat increased by the formation of interloc
crystals — occasioned by the transformation of the quartz tc
lower specific-gravity forms. vSeaver1 likewise maintained tha
interlocking crystals of well-burned silica bricks increases '
strength.

If all other properties of a silica brick made from quar
are found satisfactory, it is not usually found that the brick
fail in use because of lack of strength. The Baraboo and Me
varieties (Table 5) have both been thoroughly tried out in
furnace and coke oven work and both give consistently sati
tory results. The great variations in strength of the M(
brick, as compared with the Baraboo, may be due to an un<
development of interlocking crystals in the individual s
mens of the former.

In Table 6 is presented the data of crushing tests made
"Star" silica brick at various temperatures as reporter

Table 6. — Crushing Strength of Silica Brick at Various Tempi)
ruREs (Le Chatelier).

Crushing

strength of hot brick.

Temperature ° C.

Kg./sq. cm.

Lbs./sq. in.

15

170

2418

520

158

2247

670

I50

2133

800

139

1977

950

125

1778

1050

120

1707

I200

85

1209

1320

62

882

1460

50

7"

I540

37

526

1600

30

42 7

17OO

12

1712

1 K. Seaver, "Manufacture and Tests of Silica Brick for the By-P
Coke Oven," Trans. Am. Inst. Mining Eng., 53, 125-139 (1916).

2 Extrapolated. .

AMERICAN CERAMIC SOCIETY. 493

eChatelier.1 The 17000 C figure has been extrapolated by a
)ntinuation of the curve. This curve would seem to indicate
lat hot crushing tests, to be of the most value, should be made
: a temperature that corresponds to the working conditions —
ich as the temperatures of the crowns of steel furnaces. Orr

made at any other temperature, the strength obtained should
) compared to that at the corresponding temperatures on pre -
iously prepared curves of the above nature. Such prepared
irves should of course include varieties of materials similar
) those under test.

McDowell,1 in his tests of silica brick which had been burned
om one to ten times in a commercial kiln, found that at the end
j the first few burns the strength had reached a maximum —
iter which it declined slightly. His theory is that this de-
■ease is due to a slight rupture of the brick — caused by re-
eated heating and cooling — rather than to an inherent weak-
ess of one or both of the low specific-gravity forms of silica.

In making load tests on silica brick, it was soon found that,

heated to 700 ° C at the same rate as clay brick (2700 C in 20
linutes, 5200 C in 40 minutes, and 6700 C in 60 minutes), the
)ecimens invariably spalled; and that spalling apparently
:ased when the rate of heating below 5000 C was decreased to
d° C in 15 minutes. Measurements of the linear expansion of
ricks during the load tests indicate that the spalling is prac-
cally all due to the alpha-beta cristobalite inversion — which
ikes place at from 220-275 ° C.2 In Table 7 is given the data
r such a test. From this data it is seen that there is a decided
cpansion at the lower temperatures. This is followed by a
'adual but very slight expansion until a temperature of approxi-
lately 14000 C is reached — at which point a rather decided in-
case in the rate of expansion is noted. This expansion, which
icomes apparent at 14000, is approximately equivalent to the
^rmanent expansion of the brick. Or, in other words, this ex-

1 hoc. at.

2 The temperature depends upon the previous heat tieatment which the
aterial has received.

494

JOURNAL OF THE

I Aiii.u 7.— Expansion ok Brick No. 27 under a Load of 25 POUNDS im:k

Square Inch.

Time in hours. Temperature ° C. Per cent linear expansion.

O.OO 19 O.O

I .OO 200 O.O

2.50 500 O.34O

3-00 650 O.363

4.5O I200 O.403

4-75 1250 0.40S

5.00 1300 0.431

5-2.5 1330 0.431

5 .50 1360 0.657

6.00 1400 0.657

6.50 1400 0.657

7.00 1400 0.703

7.50 1400 0.840

Measured cold after test o . 158

Total expansion o .840 Per cent

Permanent expansion o . 153 Per cent

Difference compared to 0.657 at 13600 C 0.682 Per cent

pansion may be considered as due to the actual transformation
of silica from the high to low specific-gravity forms.

~-Heai~tnq Coil

Fig. 6.

-Extenseometer, designed to measure the expansion of silica brick
between atmospheric temperatures and 300 ° C.

AMERICAN CERAMIC SOCIETY. 495

In Fig. 6 is shown an apparatus for measuring the expansion
silica brick up to 300 ° C. It consists of a lever arm to register
pansion and an attached mirror for magnifying the readings
any desired extent by means of a telescope and graduated
ale. The test specimen is immersed in a bath of "Crisco,"
lich is heated electrically by means of a nichrome grid.
le temperature throughout the bath is kept uniform by
irring the liquid with a paddle. The supports between the
ecimen and bed plate are of quartz glass, as is also the rod
:tween the specimen and the lever arm. It would be of in-
rest to determine the exact temperatures at which the chief
pansions take place for bricks made from each of the leading
irieties of quartzite, and also the time required for a standard
inch brick to reach its maximum expansion when the tempera-
re is kept constant.

Load Test.

The load test on silica brick is conducted in the usual form of
ad test furnace.1 The load is 25 pounds per square inch,
lie heating rate is 50 ° C in 15 minutes from atmospheric tem-
:rature to 5000 C, 75° C in 15 minutes from 5000 C to 8oo° C,
10 ° C in 15 minutes from 8oo° C to 1200° C, 500 C in 15 minutes
am 12000 C to 13500 C and 300 C in 15 minutes from 13500 C
15000 C. A temperature of 15000 C is maintained for i1/^
turs — after which the firing is discontinued. When cold, the
ick should not have increased in length more than two per
nt (l/i" per foot), nor decreased in length more than one per
tit (Vs" per foot). The Refractory Materials Committee of
e American Gas Institute has set one per cent permanent ex-
nsion as the limit in such tests. This one per cent limit is a
sirable one if materials which will satisfy it can be regularly
tained.

General Results.

Among the more highly metamorphosed quartzites, those which

ush largely by breaking through the original grains are apt

yield brick of lower porosity than those in which the breaking

largely through the interstitial material between the built-up

'• U. S. Bur. Standards, Tech. Paper 7.

496 JOURNAL OF THE

grains — or those in which a large proportion of the original grains
break free from the interstitial material while the more friable
materials, which crush almost entirely to individual grains, show
a very high porosity and yield brick which are weak and friable!

Innovations in the methods of manufacture may be ad visa
ble in special cases in order to obtain greater uniformity of prod-
uct. One such innovation is to control the grind so that the
percentages of the various-sized particles in the mix will remain
constant at all times and to then machine mold the brick at a
pressure of approximately 1500 pounds per square inch in order
to obtain brick of smooth finish and uniform size.

The porosity of a finished silica brick depends primarily upon
the porosity of the unburned mix. However, in so far as the
burning does affect their porosity, the lowest porosity will
probably be obtained by not allowing the temperature of the
kiln to rise above 13500 C until a large percentage of the quart?
has had time to be converted to the lower specific -gravity forms
Rapid heating of a silica brick to temperatures above 13500 C
while a large proportion of the silica is still present as quartz
invariably produces a friable (punky) brick. To obtain a bricl
of comparatively low specific gravity, however, the temperatun
should eventually be slowly raised to a maximum corresponding
to cones 18-20.

The strength of a brick of slight or medium burn (gauged b;
present practice) is probably due to the glassy bond formed b;j
the interaction of the basic fluxes with the silica, while in wel
burned brick it appears that the strength is greatly augmente
by the formation of interlocking crystals of the low specificl
gravity forms of silica. As a general rule, it has been foun
that silica brick, made from quartzite giving entire satisf actio I
in other ways, usually meet the strength requirements of se
vice.

Inspection of Ware Based on the Above Observations.

Some of the properties of a finished brick which may be used
judging its probable value in use are as follows: Chemical
composition, specific gravity, porosity, cold cross-breakii
strength, behavior in the load test, the hot crushing streng j

AMERICAN CERAMIC SOCIETY. 497

it the temperature at whieh bricks are to be used), and the
mount of interlocking crystals — determined by viewing thin
actions of the brick under the microscope — and the softening
;mperature of the brick. For the best quality brick these proper-
es, based on available data, approximate the following:

The silica content should not be much under 94 per cent,
ricks containing lower percentages of silica are less refractory,
he alkalies should not exceed 0.5 per cent. Greater percentages
:duce the refractoriness in direct proportion to their amount,
he limit of iron oxide was previously assumed to be 0.5 per
Int. However, greater percentages, up to 1.6 per cent, do not
Dpear to materially affect the refractoriness of the brick. The
tnit for lime is usually set at 2.0 per cent, although larger
Tiounts of lime (up to several per cent) do not greatly lower the
>ftening temperatures. However, in commercial practice, more
lan 2.0 per cent of lime is considered undesirable.

The average specific gravity for brick made from Medina
nartzite, as shown in Table 3, is 2.357 and, for brick made from
araboo quartzite, it is 2.399. Hence, bricks made from similar
laterials may be considered as having received a medium burn

their specific gravities correspond to these figures.

The practical limits of porosity which first quality silica brick
'e apt to have are, Medina 22.80 to 31 .52 per cent, and Bara-
do 23 . 45 to 25 . 66 per cent. However, porosities lower than
lese might be desirable.

The cold cross-breaking tests, as shown in Table 4 and from
:her data, should not show an average modulus of rupture
uch below 500 pounds per square inch. The reliability of this
jure is considerably impaired by the comparatively small amount

data upon which it is based.

After having been tested under a load of 25 pounds per square
ch at 15000 C, a brick should have neither expanded (linearly)
ore than two per cent (preferably 1 .0 per cent), nor contracted
ore than one per cent. If desired, the load test furnace may be
ranged so that the specimen can be crushed hot at the com-
etion of the test. It would seem desirable to conduct this
st at the temperature at which the bricks are to be used.

49* JOURNAL OF THE

A micro-examination of a thin section of a brick will detect
the amounts of interlocking crystals of the low specific-gravity
forms which are present, and thus indicate the degree of burning
which the brick received in manufacture.

First quality brick should not have a fusion temperature much
below that of cone 31.1 Cones of the most refractory silica
brick bend over at cone 32 V2.

Summary.

In summing up, we may say that the bulk of all silica bricks
are made from the quartzites of early geologic age, and that, as a
rule, the quartzites most suitable for the manufacture of silica
brick are those which range from the slightly porous to the highly
metamorphosed impervious rocks — having tightly interlocking
grains. These may be distinguished in thin section under th<
microscope, by the fact that the interstitial impurities (such a
hydrated iron oxide) appear as sharply defined lines betweer
the tightly interlocking quartz grains, while in material whicl
has been but slightly metamorphosed the interstitial materia
appears in translucent, cloudy areas between the harder quart
grains. Those quartzites which crush largely by a breakin
through of the original grains are apt to yield silica brick of lowe
porosity than those in which the breaking is largely throug
the interstitial material between the built up grains, or those i]
which the original grains break free from the interstitial materia

In special cases, to obtain greater uniformity of product,
may be advisable to control the grind by unusual manipulation- ■
so that the percentages of various-sized particles in the mix wi ]
remain constant at all times — and to then mold the brick t,j
machine.

The porosity of a finished silica brick depends primarily up< (
the porosity of the unburned mix. However, the heating
brick, which contain large percentages of unchanged quarll
to temperatures above i350°C causes undue expansion — resulti' ]
in a punky product. On the other hand, maintaining the tempei <

1 When ground to pass through a 60-mesh screen and made into col
similar to the standard cones used for comparison.

AMERICAN CERAMIC SOCIETY.

499

ure of the kiln at from 12500 C to 13500 C, until a large per-
entage of the quartz in the bricks has been transformed to
he low specific-gravity forms, and then slowly proceeding to
he higher temperatures, should result in a low porosity brick
if low specific gravity.

The specific gravity of a silica brick, as determined by the
yet, dry and suspended weight method, answers as a quick
aeans of determining the degree to which the quartz originally
•resent has been transformed to the lower specific-gravity forms
i silica. But to obtain a comprehensive idea of the qualities
if any variety of silica bricks, the other properties, such as porosity,
old cross-breaking strength, softening temperature, behavior
n the load test, and appearance in thin section under the micro-
cope should also be determined.

COMMUNICATED DISCUSSIONS.

R. M. Howe: Mr. Ross mentions a great many interesting
toints arising in the field of silica brick manufacture and gives
onsiderable valuable information.

An interesting point is in connection with his screen analyses
s applied to "sand rock." A typical sand-rock analysis fol-
ows:

Retained on

Per cent.
14-mesh sieve 0.73

Through

20

40

60 "

80 "
100
150
230
230-mesh sieve.

2 .02

42.55

46.30

7 .12

0.92

o.39
0.21
o .20

Total 100.44

He also mentions the rapidly growing importance of the "specific-
ravity" test. Its value is recognized at the present time to
uch an extent that one company finishes all kilns by means of
t. The draw trials are taken from the kilns at intervals and
rtien they have a specific gravity of 2.40 the firing is stopped.

5oo JOURNAL OF THE

As the kiln is held the brick undergo further transformation

tin' specific gravities of the brick when drawn being below 2.38.
Records are carefully kept, and the firing of each particular
kiln may be controlled by this means after sufficient data has
been collected.

Mr. Ross, however, makes some statements regarding chemical
analysis which are evidently not justified by the work com-
pleted to the present time. There is no evidence on record which
proves that certain impurities cannot be tolerated. Some
materials are high in one particular impurity, but are
correspondingly low in other impurities. In view of this and of
the necessary incriminating evidence, one cannot justly say as
to what percentage of each individual impurity may or may
not be tolerated.

As a rule, however, the total percentages of impurities in good
grades of silica brick are similar. These generally run between
4 . o and 6 . 5 per cent. If, then, the total of silica averages in the
vicinity of 94.0 per cent or more, about all has been done which
can be done. A silica content of 94.0 per cent excludes any
superabundance of impurity and yet does not unjustly discriminate
against any particular combination of the same. It appears
that suitable material can be secured on a silica basis alone.
In so doing, the usual exceptions encountered when specifying
ceramic materials on the analytical basis will be avoided.

The tests recommended by Mr. Ross have the same general
aim. They all tend to develop a well-made, well-burned brick.
It is surprising to note how well these same features can be deter-
mined instantly by one familiar with a particular brand of silica
brick. Two bricks when tapped together show by their ring
nearly as much as can be learned by tests. This may be explained
by the fact that the raw material is practically the same — the
lime bond being kept within narrow limits and most silica brick
being sufficiently refractory. The "ring" test determines the
variable qualities — the burn and structure. The specific-gravity
test, however, goes beyond the scope of the "ring" test in that
it is accurate in discriminating between different brands. Some
raw materials burn to form a product of good "ring" at low

AMERICAN CERAMIC SOCIETY. 501

emperatures. Such materials are quickly detected because
f their high specific gravities. This is found to be particularly
rue when checked by the re-heating (expansion) test.

D. W. Ross: With reference to Mr. Howe's remarks on
he above paper, it is of interest to note that in his screen analysis,
1 determining the size of grain in sandrock (presumably either
lomewood sandstone or Oriskany sandstone) large percentages
rere retained on the 40- and 60-mesh sieves. This ap-
arently checks the results of our own screen analyses.

It is gratifying to know that in so short a time the specific -
ravity test for silica brick has been applied in so many useful
rays, not the least of which is the method for control of the
urning as set forth by Mr. Howe.

Mr. Howe strongly emphasizes the point brought out in the
aper, that silica brick, as manufactured in the United States,
sually fail on account of their physical properties and not on
ccount of their chemical composition. In our opinion, entirely
00 much importance has, in the past, been placed on the chem-
:al composition. With this in mind, no presentation of data
nd detailed discussion has been given in reference to chemical
omposition. Instead, the merest statement of the results ob-
ained by ourselves and others, and of the usual practice as we
nderstand it, have been set forth. This has been done in order
3 place on record information which we hope may be of assis-
ance to future investigators in this complex and interesting
eld — the effects of fluxes, etc., on the properties of silica brick.

ANTIMONY OXIDE AS AN OPACIFIER IN CAST IRON

ENAMELS.

By J. B. Shaw, Alfred, N. Y.

Introduction.

The use of antimony oxide in the replacement of tin oxide as an
opacifying agent in cast iron enamels has been the subject of a
number of investigations. The subject has always been an im-
portant one, but there has never been a time when it was so vital
as at the present. The present market price of tin oxide makes
its use in enamels almost prohibitive. Eight per cent is about
the minimum amount of tin oxide that will give satisfactory re-
sults in cast iron enamels. At 72 cents per pound, this means
that the tin oxide alone in 100 lbs. of raw enamel will cost $5 . 76.
This was the cost of 100 lbs. of enamel used by the writer in 19 13.

Antimony compounds (Xeukonin) may be purchased for 16
cents per lb. and antimony oxide is relatively cheap. When it is
considered that antimony oxide is equal to tin oxide, pound for
pound, as an opacifying agent, it is readily seen that the manu-
facturers of cast iron enamel wares have a possibility of reducing
the cost of their enamels about 100 per cent by the use of anti-
mony oxide.

Staley1 has presented a very clear statement of the difficulties
and possibilities of the problem together, with the results of the
experiments he has performed. It is not the intention to enter
into a detailed discussion of the problem or to repeat what can
be readily obtained from the above-mentioned article. Much
of the data given herein confirms Staley' s conclusions — although
on some points our opinions are contradictory. This article will
have served its purpose if it stimulates some manufacturer to -
experiments which result in the successful use of antimony com-
pounds in his enamels.

Antimony oxide has been used successfully as an opacifier in
1 Trans. Am. Ceram. Soc, 18, 173 (1915).

AMERICAN CERAMIC SOCIETY 503

>teel enamels for many years. There is no material difficulty at-
tending its use in the frit for steel enamels, the color being a
ninor consideration as compared with cast iron enamels — the
Dther constituents of the steel enamel being such as to give better
:olor than those of the cast iron enamels. It may be said that
there are no great difficulties attending the use of antimony
>xide in cast iron enamels, aside from the obtaining of satisfac-
:ory colors. The color, however, is of the first importance — -so
jreat in fact that even now, after years of experimenting and when
:he desirability for its use is so great, antimony oxide is used
mly in very limited amounts. However, its use is bound to be
extended with an increase in the knowledge of how to modify
formulas so as to secure satisfactory results.

The results described were obtained in an effort to outline
satisfactory working formulas — having antimony oxide as the
:hief opacifying agent — for commercial use. All have reference
to cast iron enamels.

Experimental.

The work was carried out in three stages as follows:

(1) Preliminary batches, hundreds of which were made, were
melted in 500-gram quantities in a 'aboratory furnace. Den-
ver fire clay crucibles were found quite satisfactory for this pur-
pose. It was generally necessary to insert the cold crucible with
ts batch into the furnace at a temperature of from 1 ioo° to 13000
2. During all of the experiments it was very seldom that the
ire-clay crucibles fai'ed under the treatment.

(2) After making laboratory trials with the enamels, those which
jave promise of satisfactory results were melted in batches of
ibout 75 lbs. The melted batches were ground and applied on
commercial wares. The preliminary melts gave a good indica-
tion as to the results which were likely to be obtained with full-
sized batches — a great amount of time, money and material be-
ing thereby saved.

(3) The enamels were finally melted in batches of about 2000
pounds and applied to commercial ware.

504 I' >URNAL OF THE

The materials used in compounding the enamels were all of a
good commercial grade, the formulas given being figured to a
basis of chemically pure materials (feldspar included). A Con
necticut feldspar having the following analysis was used:

Chemical analysis Per cent

K20 730

Na20 2 .90 Formula

AI2O3 18.60 0.62 K20 I .,_ /

cv/-» o ivt rt '4 AI0O3 < 9 4 SlOj

vSi02 70 .40 o 38 Na20 J

Fe2Oa . . 014

CaO 0.65 Formula weight, 790.0.

MgO trace

Loss 0.26

100.25

Series i.

Knowing the efficiency of cryolite as a flux and opacifier in
enamels, a number of trials were made with a view to determining
the possibility of using it in conjunction with antimony oxide.
To this end the following series of enamels was prepared:

SERi

ES I.

—Batch Weights.

No.

Feld-
spar.

Flint.

Borax

Soda
ash.

Niter

Fluor-
spar.

Cryo-
lite.

Barium
carbon-
ate.

Zinc
oxide.

Anti-
mony
oxide.

PbO

I

53

O

58

21

26

8

42

O O

O.O

2-5

0.0

2

73

O

58

28

26

8

29

O O

O.O

25

0.0

3

84

O

58

32

26

8

21

O O

O.O

2-5

O.O

4

56

24

58

21

25

8

42

O O

O.O

2 -5

0.0

5

73

24

58

28

26

8

29

O O

O O

2 5

0.0

6

84

24

58

32

26

8

21

O O

O O

25

O .0

7

56

54

58

21

26

8

42

O O

O.O

25

0.0

8

73

54

58

28

26

8

29

O.O

O O

2-5

0.0

9

84

54

58

32

26

8

21

O O

O O

2 5

O .0

10

56

0

58

I I

26

16

42

O O

O O

25

O.O

1 1

73

0

58

17

26

16

29

O.O

O.O

2-5

0.0

12

84

0

58

21

26

16

21

O.O

O.O

2 -5

O.O

13

56

24

58

II

26

16

42

O.O

O.O

2-5

O .0

14

73

24

58

17

26

16

29

O.O

O.O

2-5

0.0

15

84

24

58

21

26

16

21

O.O

O.O

2-5

0 0

16

56

54

58

I I

26

16

42

O.O

O.O

2-5

0.0

17

73

54

58

17

26

16

29

O.O

O.O

2-5

0 0

AMERICAN CERAMIC SOCIETY. 505

Series

1.—

-Batch Wi

UGHTS

— (Com

lit

■tied).

Feld-
spar.

Flint

. Borax.

Soda
ash.

Niter.

Fluor
spar.

• Cryo-
lite.

Barium
carbon-
ate.

Zinc
oxide.

Anti-
mony
oxide.

PbO.

84

54

53

21

26

16

21

O O

O.O

2-5

O.O

56

0

53

O

26

23

42

O.O

O .O

2-5

0.0

73

0

58

6

26

23

29

O.O

O.O

25

O.O

84

0

58

1 I

26

23

21

O O

O O

2-5

0.0

56

24

58

O

26

23

42

O O

O.O

2-5

0.0

73

24

58

6

26

23

29

O.O

O O

2-5

O.O

84

24

58

1 1

26

23

21

O.O

O O

2.5

0.0

56

54

58

0

26

23

42

O.O

O O

2.5

O.O

73

54

58

6

26

23

29

O O

O O

2 -5

0.0

84

54

58

1 1

26

23

21

O .O

O .O

2-5

0.0

no

0

58

0

24

23

29

O .O

O .O

30

O .0

73

25

76

5

19

23

29

O O

O .O

30

0.0

73

43

95

0

19

23

29

O.O

O .O

3.0

0 .0

73

43

58

1 1

19

16

29

20.0

O O

30

0.0

73

43

58

11

19

16

29

O.O

8.0

30

0 .0

73

43

58

0

19

16

29

20.0

8.0

30

0 .0

73

43

67

8

2

12

29

20.0

12 .O

50

O.O

73

43

76

6

2

12

29

20.0

12 .O

5 0

0 .0

73

43

86

3

2

12

29

20.0

12 .O

50

O .0

73

43

95

0

2

12

29

20.0

12 .O

50

O.O

56

66

57

0

7

8

29

20.0

l6 .O

50

22 O

56

66

57

0

7

8

29

20.0

20.0

50

I I O

56

66

57

0

7

3

29

O .O

28.O

50

II O

90

0

57

15

17

8

O

O O

24.O

8.0

I I O

Series i.—

-Formulas.

KNaO

CaO.

Ba

0.

ZnO.

AhOs.

Si(32.

B2O3.

PbO.

0.9

O

. 10

O

.0

O.O

0

20

O

.60

0.30

O.O

0.9

O

. 10

O

,0

O.O

0

20

O

.60

0.30

O.O

0.9

O

. 10

O

,0

O.O

0

20

O

.60

0.30

O.O

0.9

O

. 10

O

,0

O.O

0

20

I

OO

0.30

O.O

0.9

O

. 10

O

0

O O

0

20

I

OO

0.30

O.O

0.9

O

. 10

O

0

O.O

0

20

I

OO

0.30

O.O

0.9

O

. 10

O

.0

O.O

0

20

I

■50

0.30

O.O

0.9

O

. 10

O

,0

O .O

0

.20

I

■50

0.30

O O

0.9

O

. 10

O

0

O.O

0

20

I

50

0.30

O.O

0.8

O

.20

O

0

O.O

0

20

O

.60

0.30

O.O

0.8

O

.20

O

0

O.O

0

20

O

.60

0.30

O.O

0.8

O

.20

O

.0

O O

0

20

O

.60

0.30

O.O

0.8

O

.20

O

.0

O .0

0

20

I

.OO

0.30

O.O

0.8

O

.20

O

.0

O.O

0

20

I

OO

0.30

O.O

5o6 JOURNAL OF THE

Series i. — 1

Okmci.as

(C'o)ili

lined).

'5

0.8

0.20

0.0

0.0

0.20

1 .00

0.30

0 0

16

0.8

0.20

0.0

0 0

0.20

1 .50

0.30

0 <>

'7

0.8

0.20

0.0

0.0

0.20

1 .50

0.30

0 .0

iS

0.8

0.20

0.0

0.0

0.20

150

0.30

0.0

19

0.7

0.30

0.0

0.0

0.20

0.60

0.30

0.0

20

0.7

0.30

0.0

0.0

0.20

0.60

0.30

0.0

21

0.7

0.30

0.0

0.0

0.20

0.60

0.30

0.0

22

0.7

0.30

0.0

0.0

0.20

1 .00

0.30

0 0

23

0.7

0.30

0 .0

0.0

0.20

1 .00

0.30

0 .0

24

0.7

0.30

0.0

0.0

0.20

1 .00

0.30

0.0

25

0.7

0.30

0.0

0.0

0.20

150

0.30

0.0

26

0.7

0.30

0.0

0.0

0.20

1 .50

0.30

0.0

27

0.7

0.30

0.0

0.0

0.20

1 50

0.30

0 .0

28

0.7

0.30

0.0

0.0

0.20

1 .20

0.30

0.0

29

0.7

0.30

0.0

0.0

0.27

1 .20

0.40

0.0

30

0.7

0.30

0.0

0.0

0.20

1.50

0.50

0.0

31

0.7

0.20

O.I

0.0

0.20

1.50

0.30

00

32

0.7

0.20

0.0

0 .10

0.20

1 50

0.30

0.0

33

0.6

0.20

0. 1

0 .10

0.20

1 .50

0.30

0.0

34

0.6

0.15

O.I

0.15

0.20

1.50

o.35

0.0

35

0.6

0.15

O.I

0.15

0.20

1 .50

0.40

0.0

36

0.6

0.15

O.I

0.15

0.20

1 .50

o.45

0.0

37

0.6

0.15

0. 1

0.15

0.20

1 SO

0.50

0.0

4i

O.I

0.40

0. 1

0.20

0.17

1 .70

0.30

0. 10

42

0. 1

0.40

O.I

0.25

0.17

1.70

0.30

0.05

43

0. 1

0.40

0.0

0.35

0.17

1 .70

0.30

0.05

44

0. 1

0.39

0.0

0.30

0.16

0.96

0.30

005

Results — Series 1.

Enamels Nos. 1 to 27 were first tested on small trials. All of
these are very fusible with the exception of Nos. 25, 26 and 27.
The most brilliant enamels were those highest in silica and low-
est in cryolite. Very dull (matt) enamels resulted from high
cryolite and low silica contents. The color was fairly good
throughout the series but those having the highest gloss — highest
silica — generally had the best white color.

The solubility of all of the enamels was determined by partly
immersing in water and subjecting them to 15 pounds' steam
pressure for 2 hours in an autoclave. The results were as follows:

AMERICAN CERAMIC SOCIETY. 507

Nos. 1 to 6 — Very badly decomposed.

Nos. 7 to 8 — Much less than preceding.

Nos. 7 to 9 — Hardly noticeably affected.

Nos. 10 to 12 — Badly etched.

Nos. 13 to 15 — Very slightly affected.

No. 16 — distinctly etched.

Nos. 17 to 27 — Unaffected.

The above results may be interpreted as follows:

A. To increase the fusibility: (1) increase alkalies, (2) increase cryolite,

(3) decrease silica.

B. To increase lustre: (1) decrease cryolite, (2) increase silica.

C. To increase opacity: increase the content of cryolite and antimony.

D. To decrease solubility: (1) increase fluorspar,1 (2) decrease cryolite,

(3) increase silica.

E. To purify color: (1) decrease cryolite, (2) increase silica.

After studying the enamels of the series up to No. 27, 75-pound
batches of Nos. 23 and 28 were made and applied on commercial
ware. The color and opacity were good but both crazed very
badly— showing the influence of cryolite in producing crazing.
The silica content of these enamels is much higher than that of
good commercial enamels not containing cryolite.

Enamels Nos. 29 to 33, inclusive, respresent the attempts
made to cure the crazing in No. 28. These are all fine white enam-
els having a high luster. When tested for solubility, No. 32
showed no mark and No. 33 was only slightly dulled in a 2-hour
autoclave test.

Nos. 29, 30 and 31 are somewhat decomposed but decidedly
superior to Nos. 23 and 28. Nos. 32 and ^3 were both too hard
and lift or craze in the curves when applied on commercial ware.
Nos. 34 to 37 were made in an attempt to cure crazing by an in-
crease of the boric acid content. The crazing gradually de-
creased up to No. 36, which fitted the iron but was decidedly
lacking in covering power.

It was therefore decided that in order to retain the cheapness
and stability of the enamel, crazing must be corrected by increas-
ing the silica content. This was attempted, as is represented
by enamels Nos. 41 to 43, which crazed and were decidedly
harder (more infusible) than the ordinary cast iron enamels. In
1 This holds true only when smelting is very complete.

S()S

JOURNAL OF THE

order to get a sufficient thickness of enamel to give a good white
coat, three or more dredgings were necessary. This required a
departure from the usual factory practice and the results were
not sufficiently promising to justify their adoption.

Series 2.

The results obtained with the cryolite enamels led to the
conclusion that, while good colors were easily obtained, it would
not be possible to utilize the fluxing power of cryolite to advan-
tage and that the range of composition within which they would
fit the iron would be small. In other words, the factor of safety
in cryolite enamels is low. Because of this fact it was deemed
advisable to dispense with the use of cryolite and develop satis-
factory formulas without it.

To this end the following series was prepared:

Series

2.-

-Batch Weights

No.

Feldspar

Borax

. Soda ash.

Niter.

B

car

arium
bonate.

Zinc
oxide.

Antimony
oxide. I.itli.i

47

78

57

49

7

O

8

20

48

78

57

39

17

O

16

20

49

78

57

29

n

O

20

20

50

78

57

49

17

20

0

20

5i

78

57

39

17

40

0

20

52

78

57

29

17

60

0

20

53

78

57

49

n

IO

4

20

54

78

57

39

17

20

8

20

55

78

57

29

[7

30

12

20

56

78

57

22

[7

30

16

20

57

78

57

17

[7

30

20

20

FORMULAS.

No.

KjO.

NaiO.

BaO.

ZnO.

A1203.

Si02.

B2O3.

Sb2Oa.

PbO.

47

0. 14

0.71

O

0

O

IO

O

'4

O

84

0.3

0.07

O.05

48

0. 14

0.61

O

0

O

20

O

14

O

84

0.3

0.07

O.05

49

0. 14

051

O

0

O

30

O

14

O

84

0.3

O.O7

O.05

50

O. 14

0.71

O

10

O

O

O

14

O

84

0.3

O.O7

O.05

51

O.14

0.61

O

20

O

O

O

14

O

84

0.3

O.07

O.05

52

O.14

0.51

O

30

O

O

O

14

O

84

0.3

O.O7

O.05

53

O.14

0.71

O

05

O

05

O

14

O

84

0.3

O.O7

O.05

54

O.14

0.61

O

10

O

IO

O

H

O

84

0.3

O.O7

O.05

55

0. 14

0 51

O

15

O

15

O

14

O

84

0.3

O.O7

O.05

56

0. 14

0.46

O

15

O

.20

O

14

O

84

03

0.07

O.05

57

0. 14

0.41

O

15

O

25

O

14

O

84

03

0.07

O.05

AMERICAN CERAMIC SOCIETY.

5«9

Results -Series 2.

This series (Nos. 47 to 57) was a total failure. It is given here
in order to bring out the one point of interest which it revealed.
All preceding and subsequent enamels in this investigation were
first melted in small crucible batches. These 10 enamels were
prepared as usual and placed in the furnace, but not one of them
could be melted down to a homogeneous solid enamel. All
frothed and boiled vigorously and retained the consistency of
soap suds after intense heat treatment. A repetition of the series
gave the same results. A satisfactory explanation of this phe-
nomenon has not been found. It will be noted that the enamels
contained no lime and it was found that a very slight addition of
either fluorspar or whiting would cause the enamels to fuse down
normally. It has been noted in many cases, since this observation
was made, that enamels containing no 1 me or cryolite are likely
to be frothy and smelt with great difficulty and are also likely
to repeat this boiling when burned on the ware.

Observations made on this point in many instances substantiate
the statement that no very satisfactory enamel can be made in
the absence of lime (fluorspar or calcium carbonate), lead or
cryolite.

Series 3.

The enamels of the following series represent the type of enamel
which has been found by the writer to give the best results when
antimony oxide is used as the opacifier:

.Series 3. — Batch Weights.

No.

58
59
60
61
62

63

64

65
66

Feldspar. Borax.
78 57

78
78
78
90

95
100

84
90

57
57
57
57
57
57
57
57

Soda Lith- Barium Zinc Antimony

ash. Niter. Fluorspar, arge. carbonate, oxide, oxide.

39 17 8

29 17 8

19 17 8

6 17 8

15 17 8

15 17 8

13 17 8

17 17 8

15 17 8

0

8

17

0

16

17

20

16

17

40

16

17

0

24

17

0

24

17

0

24

17

10

24

17

10

24

17

5io JOURNAL OF THE

Formulas.

No. K?(> NfuO. C»0. BaO. ZnO. PbO. AJ1O1. SiOi. MjOj.

58 0.14 061 0.10 0.0 0.10 0.05 0.14 084 0.3

59 0.14 0.51 0.10 O.O 0.20 0.05 O.14 0.84 0.3

60 0.14 0.41 0.10 0.10 0.20 0.05 0.14 0.84 0.3

61 0.14 0.31 0.10 0.20 0.20 0.05 0.14 0.84 0.3

62 0.16 0.39 0.10 0.0 0.30 0.05 0.16 0.96 0.3

63 0.17 0.38 0.10 0.0 0.30 0.05 0.17 1.02 0.3

64 0.18 0.37 0.10 0.05 0.30 0.05 0.18 1.08 03

65 0.15 o 40 o 10 0.05 0.30 0.0 0.15 0.90 o 3

66 0.16 0.39 0.10 0.10 0.30 0.0 O.16 0.96 0.3

The control of crazing and shivering is best accomplished by
varying the silica content. The fusibility of these enamels is
quite similar to that of the ordinary tin enamels and the opacity
is quite as good as that of enamels in which an equal amount of
tin oxide is used.

The color depends largely on the purity and color of the anti-
mony oxide used. Unless the very purest white oxide is used,
the color of the enamels is liable to be similar to that of the oxide
itself.

Excessive smelting will produce a bluish green color in any
antimony enamel containing fluorspar and a low percentage of lead.

A high lead content is not permissible with antimony because the
combination produces a yellow color. Blue-white can be con-
verted to pinkish or yellow-white by the careful addition of
Venetian Red or Prince's Metallic Brown (ferric oxide) up to
0.4 per cent of the weight of the raw batch.

Thoroughly oxidizing conditions (plenty of niter) must be
maintained at all times during the smelting of antimony enamels.
Reduction causes the development of a green color. Manganese
dioxide (up to o . 4 per cent) may be used to mask the blue-green
color of the fluorspar-antimony compound. An ivory-green color
results from the use of low fluorspar or insufficient smelting or
both. The addition of 1.0 per cent of lead oxide will some-
times convert a useless leadles.s antimony enamel to one having
a pleasing cream color.

Conclusion.

In order to produce antimony enamels of satisfactory color
the following should be emphasized:

AMERICAN CERAMIC SOCIETY. 511

1. Use only the purest, white, antimony oxide.

2. Use a little fluorspar (not over 5 per cent of the raw batch).

3. Be sure to maintain oxidizing conditions during smelting
(use plenty of niter).

4. Control the color by the addition of ferric oxide, manganese
dioxide and lead oxide.

5. Cryolite is unsatisfactory as a flux or opacifier in anti-
mony enamels but small quantities (up to 3 per cent) will aid
in producing a good color and without introducing any trouble-
some element.

6. Extreme care in proportioning the raw materials, very
careful and thorough mixing, and proper smelting, will insure
the successful use of antimony oxide.

Alfred University,
Alfred. N. Y.

DISCUSSION.

E. P. PosTE: Mr. Shaw's paper brings to mind the work of
vStaley on the control of crazing in cast iron enamels, in which he
mentioned the theoretical consideration of the coefficient of ex-
pansion. I wonder if Mr. Shaw, or anyone else, has had ex-
perience in a case such as he mentioned, i. e., of merely increasing
the silica content in order to slightly modify the entire formula
so that the theoretical coefficient of expansion will remain the
same and in such a way as to retain the same fusibility and at the
same time retain the same theoretical coefficient of expansion.

Mr. Staley: I did not understand the remark that was made
with reference to my paper.

Mr. Shaw: I made the statement that cryolite is not detri-
mental to the color of antimony enamels; that fluorine, intro-
duced as cryolite, does not produce an unsatisfactory color, and
that this is in contradiction to the opinions expressed by Mr.
Staley and published in a preceding volume of the Transactions.

Mr. Staley: There seems to be some misunderstanding of my
statement. I stated that the objectionable blue color of some
antimony oxide enamels is due to a combination of fluorine,
boric oxide, calcium and antimony, and that if the calcium and

5ia JOURNAL OF THE

antimony arc in certain proportions and the percentage of fluorine

increased — -whether by the addition of fluorspar or cryolite — an
objectionable blue color will result. I have used large quantities
of cryolite in oxide of antimony enamels —in fact to my knowl-
edge practically all of the recipes which arc being used com
mercially for making oxide of antimony enamels contain cryolite.
As near as I can remember, in the empirical formulas of the
oxide of antimony enamels, which I know have been used in
factory practice, the total amount of sodium and potassium
ordinarily found in American enamels falls, in nearly every case,
between 0.4 and 0.5 equivalent. The silica content of anti-
mony oxide enamels in this country varies from about 0.75 up
to possibly 1.1 equivalents. Mr. Shaw has developed a type of
enamel higher in alkalies than those commonly used in this
country.

Mr. Shaw: I would like to ask Mr. Staley: how would he
secure opacity from cryolite in an enamel containing 0.3 equiva-
lent of KNaO? if he had enough cryolite to secure any fluxing
action, how would he keep the KNaO down to 0.3 equivalent?

Mr. Staley: When the total is as low as 0.3 equivalent, the
alkali practically all comes from the feldspar and the fluorine is
introduced as fluorspar, but if the total alkalies run up to 0.4
equivalent, practically 0.1 equivalent may come from cryolite.
The cryolite in percentage weight will run around about 5 per
cent of the melted weight of the enamel. Large percentages are
not commonly used.

Mr. Shaw: I recommend in the formulas that some cryolite,
1 to 3 per cent, be used. You will not secure opacity with 5 per
cent of cryolite. If you use less than 8 to 10 per cent of the
raw weight, you do not get much fluxing or opacifying value.
Referring to the last slide on the screen. Here is shown the type
of formula which I not only have been using but am using today
on cast iron enamels. I am not stating that anyone here can take
any one of these formulas and make a cast iron enamel; that is,
at the first trial. There are a number here who can work out a
satisfactory enamel from any one of the formulas shown, but

AMERICAN CERAMIC SOCIETY. 513

they will have to correct for color, as I stated in the paper. The
color is the troublesome part. Here are shown formulas of
enamels containing from 0.45 equivalent KNaO up to 0.75
equivalent KNaO. I can take any formula shown and make a
satisfactory cast iron enamel. There is no cryolite in any of
them, but 3 per cent of cryolite may be added without any detri-
ment to the enamel. I am unable to see the benefits to be de-
rived from the use of cryolite.

.s i -4 JOURNAL OF THE

AMERICAN CERAMIC SOCIETY.

Acquisition of New Members during October, 1918.

Associate.
H. M. Thompson, Hazel Atlas Glass Co., Washington, Pa.
M. R. Scott, Bausch & Lomb Optical Co., Rochester, N Y.
James Gillinder, 11 Orange St., Port Jervis, N. Y.
Miss Dorothy P. Chapman, 31 Shattuck St , Worcester, Mass.
Sanjiro Yamada, Asahi Glass Co., Tokio, Japan.

A. L. Koch, Box 426, Barracks No. 1, Cleveland, Ohio.
Joseph Boughey, 584 Roosevelt Ave., Trenton, N. J.
Miss Anna K. Silver, 9 Hawthorne St., Worcester, Mass.
Edwin L. Hettinger, 1325 Mineral Spring Rd., Reading, Pa.

B. B. Goldsmith, 19 E- 74th St., New York City.
J. F. Sheehy, Alhambra Tile Co., Newport, Ky.

Contributing.
Vitro Manufacturing Co., Pittsburgh, Pa.
Ohio Pottery Co., Zanesville, Ohio.

i"'t>

JOURNAL

OF THE

AMERICAN CERAMIC SOCIETY

A monthly journal devoted to the arts and sciences related to
he silicate industries.

rol. 1 August, 1918 No. 8

EDITORIALS.

THE ANNUAL MEETING.

The decision of the Board of Trustees to hold the next annual
leeting in Pittsburgh on February 3rd, 4th and 5th appears to
e a logical one from every viewpoint. Pittsburgh is centrally
)cated, not only as regards railroad facilities, but particularly
s regards the distribution of the members of the American
-eramic Society. The meeting should, therefore, be unusually
rell attended as have those which have been held in Pittsburgh
r vicinity in the past.

It is to be hoped that the appeal of the Program Committee
Dr papers to be presented at this meeting will meet with a gratify -
lg response from the membership of the Society. It is to be
articularly hoped that our newer members will come to the front
nd do their share in making this meeting a banner one.

Although our membership has largely increased during the
ast two years, it will not be an easy task to produce the quantity
f high-class contributions necessary for the program of this meet-
ig. When we consider that most of our Government and Uni-
ersity laboratories have devoted a large part of their attention
3 the solution of the numerous War problems which have arisen,
nd that many of our investigators have been in the
ervice, we are convinced that unusual efforts will be necessary
d bring out the papers for this meeting. The readjustments
ecessary in many of the Ceramic industries have not been con-
ucive to careful and exhaustive researches of a peace time
ature.

5i6 JOURNAL OF THE

It behooves every member of the Society to exert himself in
soliciting and producing contributions for the program of this
meeting.

TO OUR MEMBERS.

For Americans, November eleventh, 191 8, was a far more
eventful day and a day fraught with far more radical change
in the order of everything than was April 6, 1917.

War, preparation for war and prosecution of war demanded
much of all of us, and particularly did it demand a quick adjust-
ment or sudden abandonment of manufacturing, all for a defi-
nitely known purpose, and to a definitely known result.

We are now faced by Reconstruction: What of labor, money,
prices, etc.? Will freight rates limit or prohibit sources of sup-
plies and markets for manufactured goods?

We do know that our boys are coming back, most of them im-
mediately— our transportation facilities will be required less and
less for troops and munitions. We now have better equipped
rail facilities and a growing merchant marine, all of which spells
Prosperity. None can see and only a few — a very few — dare
predict when, or how the prosperity will come. Rumors of
prosperity and of depression are deductions from the same data
and conditions, and by equally well posted minds.

These things are sure; Americans, individually and collectively,
have more money, are more optimistic, have more resources,
and have learned the value of preparedness so — Now that the
Governmental restrictions and guidance, and committee inter-
ference are no more, each individual and concern must again
gird for competition. The clarion call for Ceramic workers is to
Organize — Organize — for all purposes but in no other as much
as for strength technically.

As sure as the seas are now wide open and traversed by hundreds
of more freighters, so surely will competition be more keen. War
found us unprepared — behind the times. We caught up — but
was it not most largely because of pooled interests and federal
guidance? Will Reconstruction find us unprepared?

The American Ceramic Society has gone a long way on the
necessary road to preparedness by the several changes in rules,

AMERICAN CERAMIC SOCIETY. 517

establishment of Local Sections, publication of a Journal, but it
must go much farther. We must organize the Professional
Divisions at once and make a drive for new members. We must
have more contributions to our literature — both practical and
technical. We must have more cooperation with other scientific
societies.

Now all boost — get together — bring in your neighbors that
we, united in one for effective team work, may the better and the
quicker prepare for the technical demands of a world-wide com-
petition. The Society is yours and the officers are your servants
— write to them your ideas — contribute to the Journal — and be
sure to remember that your Society — your Organized Technical
Preparedness Effort — needs the Contributing Membership of your
firm and Associate Membership of your fellow-worker.

BOOST! for Prosperity is here.

ROSS C. PURDY,

Chairman of Membership Committee.

DIRECTORY OF DEALERS.

Attention is called to the advertisement on another page of the
'Directory of Dealers in Raw Ceramic Materials," published by
the American Ceramic Society.

This pamphlet was compiled by the Sub-Committee on Clays
ind Raw Materials for the Clay Industry, and contains the names
rf 671 firms dealing in kaolins, ball clays, flint, feldspar, plastic
md flint fire-clays, limes, etc. These are listed in two ways;
irst, the firms are placed alphabetically, with the chief ma-
:erials carried by each; second, the materials are arranged alpha-
jetically and the firms from whom they may be obtained are,
isted under each. There is a supplementary list of special glaze
md body materials, with the names of the dealers carrying them.

The bulletin is comprehensive and will be valuable to those
ivho buy the various raw materials of the ceramic industry.

ORIGINAL PAPERS AND DISCUSSIONS.

NEWCOMB POTTERY.

I. HISTORY OF NEWCOMB POTTERY.

By Mary G. .Sheerer.

The Newcomb Pottery was started through a desire of the
Art Department of Newcomb College to provide a means of
bringing the work of its graduate students before the public.
There was little or no opportunity in the South at that time
(1894-5), f°r these students except to teach, so the Department
undertook to make a way for them.

The governing motive of the Newcomb Art School has been to
bring art into every phase of life ; not to confine it to the teaching
of picture making, or statuary, but to connect it as well with
household furnishings, such as pottery, china, textile decorations,
book-binding, jewelry, etc.

In 1885 an art leagure was formed in New Orleans under the
supervision of Wm. Wordwerd, professor of art at Tulane Uni-
versity. The league conducted a pottery for five years and many
very creditable pieces came from its kilns, but, owing to lack of
capital, it was unable to continue. Like so many other efforts
which seem to fail, however, it proved to be the suggestion for,
the more substantially planned pottery which followed it some;
five years later at Newcomb.

Active experiments were begun in October, 1895. In one endl
of a large room was built a round kiln with four fire boxes?!
There were in the same room a kick-wheel, drying shelves I
tubs for clay, tables, decorator's working outfits, exhibitiorlj
shelves and tables, glaze pots and a hand-mill for grinding. Ncl
tenement room was ever used for so many purposes, but then

AMERICAN CERAMIC SOCIETY.

5i9

Fig. 1. — First home of Newcomb Pottery.

Fig. 2. — Present home of Newcomb Pottery.

520

JOURNAL OF THE

was interest, enthusiasm, and activity aplenty — and great ex-
pectations!

The services of a potter from the Golf Juan Pottery at Cannes,
France, were secured and to him was entrusted the making of
forms on the wheel, glazing and firing.

The writer came from the Cincinnati Art School to conduct
classes in the decoration of pottery and its appreciation. The
president of the College, and the director of the Art School with
the well-equipped students from the Newcomb Art School, com-
pleted the personnel of the pottery force.

A cream colored clay and a red clay, locally procured, were
washed by hand in tubs. Clear glazes, from French recipes,
were ground in a hand-mill and experiments were made on both
both clays by slip painting, biscuit painting, inlaying and
modeling. From a sample collection of underglaze colors, two
blues, a green and a yellow were selected. Considerable work
was done with slip painting, entire back-grounds being put on
by spraying with a cup atomizer.

Design was such a strong feature at the Art School that the

Fig. 3. — Corner of present salesroom.

AMERICAN CERAMIC SOCIETY. 521

iecoration was mainly simple and good from the beginning,
3ut slip painting was found difficult to control in this warm
climate where the windows and draughts are many, so it was soon
ibolished in favor of biscuit painting.

The character of the decoration came about through a natural
growth. After abandoning slip painting the designs were drawn
ind painted on the biscuit. After unsuccessful attempts to secure
:he desired texture on the clay surface as it came from the wheel,
t was decided to produce the texture with the sponge on the green
>r unfired piece. This led to designing on the wet clay and in-
cising the designs and painting after the first fire. This gave more
variety and depth to the effect and the method still continues
n use.

The last few years have seen a marked growth ofi nterest in
lie hand-built pottery. The personal touch and variety of sur-
ace which these pieces afford give added charm to the glaze
ind no decoration of the form is needed.

A matt glaze which is used on this ware has a wide range and
s good in texture. Following is its formula and also that of a
in glaze used:

Matt Glaze

0.6 PbO \ .,-«"/ o-^

0 4CaO/035Al2°3lIOoSl02

Tin Glaze

f 1 -75 SiO*
0.15 AI2O3 J 0.3 Sn02

II. TECHNICAL PRACTICE AT THE NEWCOMB POTTERY.
By Paul E. Cox.

Introduction.

In view of the fact that the Newcomb Pottery will be moved
o new quarters sometime during this year, it seems fitting that
his termination of what might be called the "second period of
he life of Newcomb Pottery" be marked by a report of the tech-
lical development to the present time.

The writer believes that Newcomb Pottery has been bettered
nd that the reputation of the pottery as an art product has been
nhanced by the application of the principles taught in the ceramic
chools, and that an art pottery is great only when it recognizes

522 JOURNAL OF THE

that technical merit is a close second to good design in decora-
tion, or good taste in choice of coloring, and that a leaky, badly
glazed and fired vase, no matter how perfect the design or how
lovely the coloring, is bad and unworthy of a place of merit.
This refers, of course, only to contemporary pottery, since there
is no excuse for anything to the contrary being made in the light
of our present knowledge.

Body Mixtures. The body in use at Newcomb Pottery eight
years ago was reported as made of a mixture of Biloxi clays Nos.
i and 2, Iuka clay, Texas kaolin, and feldspar. The Biloxi clays
were secured from a bluff close to Biloxi, Miss., and were reached
by a schooner working its way up the bayou on which the clay
was found. These clays had been used locally by flower-pot and
charcoal makers and were also rather excellent for the purposes
of pottery making, but were not easily accessible — since there
was little method in handling the overburden of the deposit
and a large amount of waste material was necessarily removed
each time clay was taken out. As a consequence, when op-
portunity offered a clay that was more easily obtained, a change
was made.

The body made from the above clay mixture was covered
with a clear glaze and fired at between cones 02 and 1 . The ware
was porous, crazed badly, and had little to recommend it except
that the body was easily worked on the wheel and in the class
rooms.

Since Newcomb Pottery is indigenous to Louisiana in decora-
tion and spirit, it was desirable that the clay be obtained in
Louisiana. Samples of clay from Covington, a point across the
river in St. Tammany Parish, close to the city of New Orleans,
proved workable and the body as now used is composed as fol-
lows: 35 per cent Covington clay, 45 per cent Paducah ball-
clay, and 20 per cent feldspar. The Paducah ball-clay was chosen
only because it gave satisfaction and was easily obtained. It
might be replaced by other light-burning clays.

The above body burns dense at between cones 3 and 5 (tht
cones used), and works well on the wheel — though a little con-
trary for large pieces — does not leak when fired even a little soft

AMERICAN CERAMIC SOCIETY. 523

and the glazes stand well without crazing. The decoration being
3Ut on by cutting away part of the clay, the mixture must be
.veil lawned and free from grit in order to secure good results.
Since a thorough sponging of the surface is made by the designer
:or the sake of the proper texture for the later painting, there
iiust be little washing away of the clay in the process. The
ibove body was developed with these objects in view and has
proven successful.

According to the writer's observations of the tests of different
:lays on the wheel by the very skillful throwers working at New-
:omb Pottery, coupled with tests by himself on the wheel, he is
nclined to question whether or not a satisfactory definition of
elasticity has ever been made. The clay in use at Newcomb
Pottery would usually be rated as a very plastic one, and yet large
pieces are made from it with difficulty because the pieces settle
after the first draft is made. On the other hand, small pieces
ire made with the utmost facility. Other clays used, one from
Pish River, Alabama, for example, are not open to the same criti-
cism, but are worthless for pottery making. Still others are en-
tirely satisfactory in all respects. Tall pieces are made on the
wheel more easily from sandy clays than from fat clays. Many
very fat clays which have been tested on the wheel could not be
thrown at all, but were otherwise of the sort that would be
considered plastic. No tests other than wheel tests were ever
made.

Preparation of Body. — Owing to the fact that the sand from the
very sandy local clays worked back from the end of the shaker
lawn and into the lawned slip cistern, and that the lawns wore
out so rapidly and at such inopportune times, shaker frames
were discarded. A trough was arranged to separate most of the
sand by means of a series of baffles and the slip was then poured
through the lawn — a single one of 150 mesh. It has been found
that better clay and more of it for a given period can be had by
this arrangement than with a machine. The Covington clay
consists of about one-third white sand, which wears the metal
lawns rapidly, so that this seemingly more primitive method is
the better one.

524 JOURNAL OF THE

From the cistern, the slip is pumped into a too gallon tank to-
which air at ioo pounds pressure is admitted. This tank is
connected to the filter press and the air under pressure forces
out the water. The writer has found that a compressed air
system for pressing clay does away with the nuisance of the over-
flow valve, presses a more uniform cake of clay without soft
centers, is easier on the sacks, and that a press of clay may be
made at night while the rest of the machinery is not working.

We have found that there is no need for agitation of the slip
in the tank if there is as much clay to keep the inert matter in
suspension as there is in this body. Any industry using natural
clays can save very largely in installation cost and subsequent
repairs by the substitution of a tight tank for the slip cistern
and by running the slip by gravity to the tank. No difficulties
from sedimentation will be met with — if reasonable care is taken
to clear the tank once a week or oftener — and better clay cakes
will come from the press. The writer has had a chance to test
out many kinds of filter-press sacks and has found that the twilled
press sack material, used in most sugar houses, is better than the
canvas used in most clay plants. It does not become clogged
so quickly and is more easily washed out and does better work
in every way.

Molding and Decorating the Ware. — All ware made at New-
comb Pottery is thrown and turned — -the same wheel being used
for turning as for the throwing.

In a leather-hard condition, the ware goes to the designer,
who sketches the design and scrapes and cuts away a part of the
clay so that the design is in slight relief. After drying at room
temperature, the ware is biscuited to cone 012, and is then painted
by the artist who designed it — the colors being underglazes sup-
plied by color dealers. The glaze coating is then put on and the
ware fired a second time at between cones 4 and 5.

Glazing. — The inside glaze coating is applied by the use of a
pump — since the ware must never be touched on the outside by
the fingers. Even the dry clay pieces must not be touched in
placing in the biscuit kiln as any finger marks will show. The
surface obtained by the sponging process is so delicate that finger

AMERICAN CERAMIC SOCIETY. 525

marks are filled in by the wash of blue with which the values
are obtained and show in a very unsightly way after the glost fire.

The outside coating of glaze is applied by dipping, the pieces
being handled by the bare bottoms and the uncolored rims.
The bottoms are then sponged free of glaze, the untouched rim
is covered with glaze by means of the air brush, and two coats
are added all over with the same air brush.

Several coatings of glaze are applied. By the use of colored
dyes, confusion in the glazing is avoided. After applying the
first coating a colored (dye) glaze coating is applied to a table full of
ware. This is followed by a coating of another color and this
provides a standard glaze coat which is free from crawling and
gives to the colors put on the values desired by the decorators.
Nearly all of the losses are due to too thick glaze coating and even
with the system outlined, mistakes in judgment occur.

The glaze now in use has the following formula:

0.10K2O ] f

0.50 CaO [0.50AI2O3J 1 .40 Si02

0.40 PbO J (

The dry glaze mixture is ground in the ball-mill with denatured
alcohol and then applied with the spray. This glaze does not
flow when applied but dries at once and can be handled rather
freely. At times, it is desirable to put a coating of colored glaze
over a piece that is not so desirable from the standpoint of color
and this easy method has saved many a dollar that would have
otherwise been lost. By the use of a glaze maturing at a lower
temperature, and by grinding in alcohol as indicated, the cooler
parts of the kiln are utilized with considerable gain.

In a general way the writer has decided that, whatever the
theory under which matt glazes are developed may be, the nature
of the body on which they are to be used must be reckoned with
and, if a glaze with a wide firing range together with freedom
from crazing is desired, the clay content must be high and silica
content low. In a general way, our experience at Newcomb
Pottery agrees with that of others as to the types of formulas
giving the best matt glazes.

It will be noted that the clay content in the above matt glaze

526 JOURNAL OF THE

is very 1 ii>^li . For a time we used calcined clay and gum of
tragacanth in the glaze. The gum proving troublesome, it was
eliminated to the advantage of the product. The calcination of
the clay was next discontinued and the glaze was found to adhere
to the ware better. Difficulties from crawling were overcome by
the use of the air brush and a thin first dipping. It is the writer's
opinion that there is no reason for the crawling of art pottery
glazes, no matter what formula is used, provided the glaze work is
given careful attention.

Kilns. — The two kilns are round, down draft, one being 6
feet and the other 5 feet diameter, outside, the larger one taking
six and the smaller three bungs of saggers. Each kiln has two
fire boxes. An iron case, 3, \e" thick, encloses the brick work of
each and, in order to make wicket building easy, a flat arch tops
the doors. The doors are built up of special sagger clay blocks
made at the pottery. The flat arches were made by supporting
the outside ends of the bricks by a plate of iron, so that only the
ends of the bricks are exposed to the high temperatures — the
bricks being driven in place tightly and with no mortar.

The smaller kiln is used for biscuit — since the biscuit can be
doubled in placing. About four-fifths of the glost kiln produces
satisfactory, well-burned ware and the space left is used in a
way which will be described later.

Distillate is used for fuel. This is stored in a thousand gallon
tank in the pottery yard, the oil being forced to the kiln burners
by pressure from a small volume blower which injects the oil into
the fire box in a stream from the central tube of the burner. The
surrounding tube of the burner furnishes air in a volume sufficient
to break the oil into a spray and to furnish much of the air for
combustion. The volume blower forces air into the oil storage
tank and the air-tube of the burner is connected to a pipe that
takes its supply from the storage tank so that the oil tank serves
as an air reservoir as well. Pressure of about eight pounds is
maintained by the blower. The oil gives little trouble — aside
from the deposit of masses of carbon which sometimes divert
the path of the flame so that the ware suffers from reduction and
blistering. This does not often happen and the losses in the kiln

AMERICAN CERAMIC SOCIETY. 527

last year were under three per cent — mostly caused by other
reasons than kiln accidents. To burn the two kilns, with an
average of 120 pieces in each, requires 150 gallons of distillate
costing from 61/% to 7V2 cents per gallon delivered.

Saggers. — For sagger-grog, waste fire brick are crushed in a
laboratory crusher. The mixture of raw sagger clay is prepared
by washing in a blunger. It is washed as thick as it can be
handled in the machine and then poured into a galvanized iron
pan resting over a furnace. A coarse screen takes out the bits of
:1am shell and oyster shell, that cause "poppers" if not removed,
and the dry grog is then added. The sagger mixture is as fol-
lows:

30 per cent by volume New Orleans Clay
30 per cent by volume Biloxi No. 1 Clay
40 per cent by volume grog

The mixture is ground as stiff as possible in a wet-pan and then
wedged by hand.

The New Orleans clay is that deposited by the river and its
sand content varies. In the clay state, the saggers stand a great
deal of rough handling without damage, due to this gummy
New Orleans clay which, when fired, vitrifies so that the saggers
are both open and strong. The clay does not burn dense by itself
because of its content of sand. Saggers made of the above mix-
ture have been in use from ten to twenty times a year for five
years and, in some cases, six years, and are still in excellent
condition. The writer feels that this is a better sagger body than
is usually met with in this sort of work and under the conditions
laid down.

Placing. — The ware is burned in a low biscuit and a high glost
fire so that most of the shrinkage takes place in the second fire.
Consequently, it is not safe to place the glost pieces on stilts.
The glaze is removed from the bottoms of the pieces and — ex-
cept in the case of pieces with wide bottoms which are not liable
to topple over — the pieces are placed on the bottoms of the saggers.
Since this is done, the saggers are not glazed inside, as we have
found that in a 14-hour burn there is no advantage in so doing,
and, furthermore, the glazed saggers have a shorter life. The

528 AMERICAN CERAMIC SOCIETY.

bottom of each sagger is coated with a wash of refractory clay

and, since the bottoms of the pieces have been cleaned of glaze,
little of this adheres to the pieces. This method has almost
entirely prevented kiln losses through sticking. It is desirable
to have the bottom of a pot glazed, but not desirable enough to
risk the loss of a piece after so much work has been done upon it.

i toe of our sources of loss has been the blistering of pieces ex-
posed to a slight reducing flame and, since the fire boxes are a
large part of these small kilns, the saggers are subject to more
reduction than they would be in a larger kiln. In case the glaze
coating is too thin, the addition of another thin coating will
make the piece a choice one. In fact, the best ware is often that
which has had a third firing with a thin added coat of glaze.
However, there is comparatively little of this sort of work to be
done, so that if the piece is valuable the expense is justified.

The work reported in this paper would have been done by
any man trained for the job and there is nothing startling
to report. However, some others, with the same problems to
solve, may find helpful hints. No claim for perfection in methods
is made by the writer — this paper merely tells what has been
done. Most of the methods followed were make-shift ones, but
the Pottery speaks for itself.

Newcomb Pottgrv,
New Orleans, La.

A CONVEYOR STOVE ROOM.

By C. E. Jackson, Wheeling, W. Va.

Much progress has been made in the pottery industry within
the last few years and especially in kiln construction. However,
very little attention has been paid to one of the most vital de-
partments of a pottery, viz., the clay shop in which all of the ware
is formed. We do not believe it possible to change its fundamental
or basic operations. We assume that it will always be necessary
to use molds made of plaster of paris or of some other material
that will give the same result. Therefore, any advancement
that can be made in this department will have to be in the simpli-
fication of its well-known operations. It was with this object
in view that the writer undertook to eliminate some of the trouble
that attends the manufacture of goods, generally known as flat-
wares, and which are made on the machines called jiggers.

It has been the custom to consider the operating jigger unit
as consisting of a jigger-man, a batter-out, one or two mold
runners, and a stationary stove room. The function of this
stove room is to receive the molds on which the ware is formed
in the wet state and to give sufficient capacity to retain the same
until dry. This means that the size of the stationary stove
room is largely governed by the number of molds necessary for
the day's operations. Let us assume that this number averages
from 175 to 200 dozen per jigger-man — and that the number
of stoves in the jiggering department is in proportion to the
number of jigger-men.

Since the operating unit consists of a jigger-man and two
(or in some cases three) male helpers, we will consider an operating
unit as consisting of a jigger-man and his attendant labor. The
attendant labor has always been the obstacle in the steady pro-
duction of the jiggering department. It has usually consisted
of boys and young men, in most cases irresponsible. When this
attendant labor fails to come to work, it means that the output
for that operating unit is lost for the day, and this becomes a

530 JOURNAL OF THE

constant occurrence. Furthermore, the efforts of the jigger
man, instead of being devoted to the making of the pieces of

ware at hand, are largely given to the control of his attendant
labor. Owing to the passage of the Child Labor Law in the
various states, it became very difficult to obtain this attendant
labor, and even more so after we entered the War. Various
plans have been suggested and talked over for the elimination
of this attendant labor. In Kurope, the problem had been partly
solved by the use of a bat-making machine, but there still re-
mained the question of getting the molds to and from the stove
room which necessitated attendant help for the jigger-man.
We endeavored, therefore, to take advantage of the partially
worked out system in vogue in Europe, and, by the use of a con-
veyor operating in conjunction with a bat-making machine,
to totally eliminate all attendant labor. The writer is pleased
to state that this has been accomplished in a most satisfactory
manner.

A conveyor system, consisting of two lengths of chain operating
on parallel lines on sprockets, and carrying shelves on which
are placed the molds, was designed and constructed. After filling,
the molds are placed by the jigger-man on the shelves of the
conveyor. A movement of the chain carries the shelves to the
rear of the conveyor where they are emptied by the finisher.
The empty molds are then returned to the jigger-man for
refilling.

In an installation of this kind there are many things to be con-
sidered. There is no question that the system requires more
space than the old stationary stove room, but in a different
manner. The old stationary stove room required considerable
space running the length of the shop while the conveyor requires
depth or height. By referring to the drawing it will be
be noted that the chain travels up and down nine times. We
found this travel necessary in order to secure the length of chain
required. Had the ceiling been higher, we would have made
the conveyor higher and the up and down movements fewer.

Very few potteries present greater obstacles than ours in the
establishment of this system. However, it is necessary to under-
stand the fundamental principles of conveying machinery before

AMERICAN CERAMIC SOCIETY.

53i

"T^

^=*

\&&

ear

5EK v

'J i

532 JOURNAL OF THE

installing the system. The first decision to be made- is the number
of molds the conveyor is to carry. If the number of dozens
of ware per machine per day is decided upon, and the molds are
to be filled but once a day, then sufficient chain length to allow
the required number of shelves is necessary. However, we must
bear in mind that the production per day, using the conveyor
type of stove room, will be less than by the use of the old system —
owing to the fact that the jigger-man has no attendant labor
whatever. We have found the difference to be about 25 per
cent. While this is true per day, the number of dozen pieces
of ware produced during a period of 30 days will be greater —
owing to the fact that a jigger-man will work full time. Using
the old stove room, with attendant labor, the efficiency was not
more than from 60 to 75 per cent. It is necessary, therefore,
to take this into consideration in figuring an installation.

Sufficient heating apparatus is necessary for the drying of the
molds and ware in transit. We have found the use of exhaust
steam pipes laid on the floor under the conveyor to be the most
advantageous. In considering the installation of the steam pipe
system, it is necessary to take into account the fact that, on the
overhead return of the conveyor, the molds have been emptied
by the finisher, which means that only 75 per cent of the molds
in the conveyor are actually drying the ware.

The batting-out machine is constructed on the same principle
as the spreader— which is nothing more nor less than a flattening-
out tool working on a revolving plaster whirler. The jigger-
man places a piece of clay of the desired size on the whirler (wher
stationary) and then throws the machine into gear. The actioi
of the down coming arm forming the bat is automatic with th<
revolving whirler. After the bat has been made, the arm beinj
released, returns to its original position, the whirler being agaii
stationary. This is all being done while the operator is makin:
a piece of ware. The operator, working at right angles to th
machine, lifts the mold out of the jigger-head with the left han
and places it on the conveyor shelf which stands at the level c
the jigger box. With the same movement he places anothe
empty mold for filling.

The three mechanical movements necessary, viz., that of tb

AMERICAN CERAMIC SOCIETY. 533

>atting-out machine, the operation of the jigger, and the movement
>f the conveyor, can all be controlled by one mechanical move-
nent as exemplified in the machine for this purpose placed on
he market by the Potters Equipment Co., 200 5th Ave., New
fork, N. Y. In our case, each unit is operated separately.

Although the action is very simple, the advantages to be gained
>y this system are: First, the elimination of attendant labor;
econd; continuity of output; third, the jigger-man has nothing
o detract his attention from the work at hand. By its use we
ind that the drying is more uniform, the production is more
iniform from all angles, the ware is of better quality, and there
5 less loss in our biscuit ware room. Comparing the earning
>ower of the jigger-man against the old method, we find that
nth the new system, the average earning of the jigger-man is
ligher by about 10 per cent. We are able to keep our shop much
leaner — -which is conducive to the health of the employees. While
,n equipment of this kind is somewhat expensive, we believe
hat the advantages to be gained are sufficient to justify the
nstallation.

Warwick China Co.

PREPARATION AND APPLICATION OF ENAMELS
FOR CAST IRON.1

By Homer F. Staley. Washington, D. C.

ENAMEL MAKING DEPARTMENT.

The enamel making department, usually known as the "mill

room," comprises facilities for storing, weighing and mixing raw

chemicals, furnaces for melting the enamels, enamel dryers, and

mills for grinding and sifting the enamel.

i. Storing, Weighing and Mixing the Raw Materials.

Of course the relative arrangement of these various parts can
be varied at will. A commonly used and convenient arrangement
is a three-story building with the melting furnaces at one side or
end. The crude raw materials are stored on the third floor.
They are sifted there by mechanical riddles and then fed into
chutes leading to bins on the second floor. The sifting not only
removes any nails, pieces of twine, etc., that may have gotten into
the raw materials, but also serves to break up the lumps that com-
monly form in powdered chemicals. The absence of lumps greatly
facilitates the fusion of the enamel to a homogeneous glass.

On the second floor the weighing and mixing of the batches
are carried on. In one arrangement the bins are built in a long
row with their bottoms a few feet above the floor level. A set of
scales on wheels runs on a track in front of this row of bins. The
chemicals can be scraped by a hoe as desired from the various
bins. Gravity feed of the powdered chemicals into the weighing
box has not proven successful. Part of the time they stick and
do not feed at all, and the rest of the time they come down with a
rush. This system eliminates all shoveling, but entails con-
siderable moving of the scales and partially completed mix back
and forth. In another system, the bins are arranged in a circle
around the scales. The chemicals can be scraped into the weigh-
ing box by means of a hoe and troughs, but often are simply

1 By permission of the Director, Bureau of Standards.

AMERICAN CERAMIC SOCIETY. 535

loveled in. Multiple-beam scales, weighing at least six in-
■edients with a single setting, are commonly used.

The mixing of the raw batch should be very thorough. With
nail batches this is often done by "cutting" the completed mix
ith shovels. Care should be taken to see that each shovelful
; the mix is actually turned over in each of at least three cuttings,
fter this the whole mix should be run through as fine a riddle as
Dssible. With large batches, mechanical mixers of various
rpes are employed. In general, it may be said that batch mixers
:e more satisfactory than continuous ones, and that the drum
rpe, in which the whole mix is thrown over and over, is better
lan the trough type, in which mixing blades travel through the
itch lying in a stationary container. In continuous mixers,
le portion fed into the machine first comes out first and there-
ire the batch as a whole is never blended. When dependence

placed on mixing blades in a stationary trough or a fixed con-
liner of other shape, a layer next to the shell is not mixed in
ith the rest.

A simple and efficient batch mixer of the drum type consists
: a ball-mill without any balls. A lining is not necessary. If
ssired, three or four stout bars may be inserted midway between
le circumference and axis and extending from one end of the
till to the other. These are not necessary, neither are shelves
) carry the mix up the sides of the mill. If a batch, not filling
le mill over two-thirds full, is run in a plain ball-mill, without
alls, at a proper speed for five minutes, it will be found to be very
loroughly mixed.

A convenient location for the mixer is below the floor level
i the weighing floor but above the level of the top of the melting
irnaces. If not more than three furnaces are to be fed from one
lixer and these are built close together, the material may be
limped into a stationary hopper and fed into the top of the
lrnace, by spouts. The angle of the feed spouts with the horizon-
il should be at least 60 degrees. When the material is to be
irried a greater distance, a hopper-bottomed car or lorry running
ver the tops of the furnaces is convenient. The sides of this
opper should also form an angle of at least 60 degrees with the

536 JOURNAL OF THE

horizontal. It is too hot on top of a furnace to permit of dis-
lodging materia] that sticks in a hopper or spout.

2. Melting the Enamel.

Melting Furnaces. — The furnaces used for melting enamel are
of the reverberatory type. They are similar to the ordinary
potters frit kiln but larger, some of them holding over two tons
of melted enamel. They are simply large boxes with inclined
bottoms and low arched roofs. The raw material is fed in batches
at the top, or side, and drawn off periodically from a tap-hole on
one side. Many efforts have been made to operate melting
furnaces continuously but without much success up to the pres-
ent time. The heat is obtained from a flame passing across the
kiln just beneath the roof. Various attempts have been made
to build frit kilns that would be heated from below, without
contact of the combustion gases with the frit, but no material
has been found that will transmit the heat readily and yet resist
the intense corroding action of the molten frit. In Fig. i are
shown the drawings of a frit furnace used in several plants. It
will be noticed that the arch of this furnace does not drop down
as it approaches the end away from the fire. A dropping arch is
more expensive to build than a straight one, and proper approach
of the flame to the melt can be obtained by regulation of the
height of the arch as a whole, the slope of the floor and the outlet
of the chimney. In another type of furnace, the position of the
melting chamber is reversed from that shown in these drawings,
the flame traveling across the short dimension. In this type the
extraction of heat from the flame is not quite as complete as in
the case where the flame travels lengthwise of the melting
chamber. When the furnace is used for oil or gas, part of the fire
box is bricked up and a checker-work of fire brick is placed or
top of the brick filling.

In some cases a second melting chamber for melting grounc
coat glass has been built onto the cold end or side of the ename
melting chamber, the idea being to utilize for melting the grounc
coat glass some of the heat that ordinarily goes up the chimney
These have not proven very successful, mainly because the hea
obtained is not sufficient to do much melting and supplementary

AMERICAN CERAMIC SOCIETY.

537

538

JOURNAL OF THE

burners have to be installed. When the bridge-wall between the
two melting chambers is not built hollow, the glass from one tank
works through into the other. The melting tank for ground
coat should be built like that for the cover coat but may be smaller.
For ground coat glasses that are to be drawn while viscous, the

r-fl

r

Fig. ib.

tap-hole may be left about four inches square and stopped up by
a couple of wedge brick. Sometimes ground coat glass is melted
simply to a very viscous mass in scrap sinks, placed in the hottest
part of an enameling furnace for 18 to 24 hours.

For melting enamels, natural gas, producer-gas, or oil should
be used for fuel. With coal and coke, too much dust and ash
is deposited on the melting enamel. Whatever kind of fuel is
used, the flame should be thoroughly oxidizing, that is, a blue
flame, free from smoke and incandescent carbon — which make a
yellow flame. The least fuel is used when the draft is so regulated
that the tip of the flame just reaches across the furnace. A shorter
flame means that too much air is being used, and a longer one
means that the stack instead of the enamel is being heated.

The M elting Process. — After the raw enamel batch is put into
the melting tank, it is pushed back against the walls in every
direction from the tap hole. This is done because the deepest

AMERICAN CERAMIC SOCIETY. 539

•art of the melting tank is at the tap-hole, and if the top of the
harge is made level it is difficult to melt the material deep down
lear the tap-hole. In the course of one to four hours, depending
m the size of the batch, the melting is finished. The amount of
tirring that should be done during the melting depends on the
iniformity of the mixing of the raw batch, the type of furnace,
nd the kind of enamel. The finer the grains of the materials
omposing the batch and the more uniformly they are blended,
he less the necessity for stirring. If a furnace of large capacity,
ompared to the size of the batch, is used, so that the material
an be spread out in a thin layer, stirring is not so necessary
Jsing thoroughly blended batches and large capacity furnaces,
he author seldom found it necessary to stir tin enamels. In
rder to get uniform color in antimony enamels, stirring of the
nix toward the end of the melting period must nearly always be
•racticed.

The objection to stirring enamels, aside from the small amount
f labor involved, is contamination of the melt with iron from the
tirring rods. It has been found that cheap tool steel rods give
;ss scale than rod iron. The enamel adhering to the rods should
lever be broken off into the tank for holding good enamel. If
he rods are chilled in a separate barrel of water, the dirty enamel
nay be saved and remelted with a new batch. In order to avoid
tirring of the enamel, one company, making steel tanks, goes

0 the length of melting the enamel in what is virtually a large
ocking ladle.

Tin enamels are nearly always tapped off within a few minutes
iter bubbling ceases. It was formerly held that this was neces-
ary in order to prevent the enamel from overheating and becom-
tig contaminated with metallic lead and tin. With better con-
rol over and understanding of our fires, we find that this re-
action was not due to overheating but to reducing conditions

1 the flame and combustion gases used for melting. With a
;ood clear blue flame, a tin enamel may be heated for hours
iter bubbling ceases without changing color. As explained in
.nother section, the tint of antimony enamels varies greatly with
he amount of heating they undergo. All enamels depending

540 JOURNAL OF THE

in part on fluorides for opacity, especially fluorspar, become
less opaque on continued heating.

3. Dryers.

The enamel is tapped out into a large tank into which a supply
of cold water, amply sufficient to chill it, is running and is then
taken to the dryers. One type of enamel dryer consists of a large
revolving iron drum with a fire beneath it. A certain amount of
fine iron scale is ground off this drum into each batch of enamel.
In other cases the enamel is dried in thick layers in large pans
heated by waste gases from a furnace or melting tank or by a
separate fire. The enamel must be stirred in these in order to
get it dried in a reasonable length of time and to prevent it from
sticking to the pans, which become quite hot. In still another
method the enamel is spread in thin layers in pans placed on
racks in large drying cupboards. These are generally heated by
gas. In this method no stirring is necessary and there is little
contamination of the enamel, but the amount of room devoted to
drying must be large. Rotary dryers, lined with porcelain blocks
and having projecting shelves made by inserting rows of deeper
porcelain blocks, have been built. These are heated by blowing
hot air through them. They must be made quite large compared
to their capacity, for if the blast of air traveling through them is
strong, it blows out all the fine particles of enamel.

4. Grinding Machinery.
For grinding enamel, ball mills are used almost universally. Por-
celain brick are nearly always used for lining. Formerly flint pebbles
were used for grinding, but in late years the use of hard-burned
porcelain balls has largely superseded them. Any particles
ground off the balls show up less in the enamel than similar
particles from the dark flint pebbles, and, when a good grade of
porcelain is used, the balls wear less than the pebbles. While
very hard and resistant to abrasion, the pebbles are brittle, and
are often destroyed by splitting. The balls are worn down byj
abrasion a little more rapidly than the pebbles, but seldom split.
Mills of from 250 pounds capacity per charge up to 3000 pounds
are used. The size is largely a matter of choice, some of the
largest plants using the smallest mills. Often large and small

AMERICAN CERAMIC SOCIETY. 54 r

lills are used in the same plant. Some enamel makers claim that
nth small mills there is less contamination of the enamel by
brasion of the linings and balls, but this is largely a matter of
roper regulation of the charge and speed of the mill. Large
lills require less labor for a given capacity.

In a few cases continuous ball mills, or tube mills, have been
sed. These have not proven very satisfactory, mainly because
lie output is small compared to the amount of power used and the
loney invested in the machines. The charge in a tube mill does
ot more than half fill the machine, while in ball mills, the charge
sually fills the machine to 75 per cent of its capacity. Hence,
le charge in the ball mills is more nearly balanced on the center
f gravity of the machine and at the same time is larger — re-
nting in a greater output for a given amount of power and volume
apacity of mills.

Whatever type or size is used, great care should be taken to
le. that no metal is exposed to abrasion. This is most likely to
ccur around the door frames. The resulting contamination of
le glass causes black specks, which are especially noticeable in
ntimony enamels. Care should also be taken to prevent con-
imination of the enamel by grease and dirt from pulleys, belts
rid bearings. The wear on linings and balls will be lessened and
le contamination from that source greatly reduced, if arrange-
lents are made to empty the ball mills while they are running
t a very slow speed.

Cover enamels are ground dry to a fineness such that they will
ass a 60-mesh sieve. Ground coat enamels are usually ground
ret to extreme fineness.

5. Screens.

In a few plants the enamel is not really screened in the mill
Dom, being simply shoveled through a coarse riddle to take out
imps of enamel and small pieces of mill lining and stones. It
as been found that the small amount of coarse material in the
tiamel powder seriously interferes with the speed of application
f the enamel, and all enamel is screened in the mill room at most
lants. The screen wire used is commonly about 10 meshes
ner than that used in the dredges, *'. e., the sieves used in ap-
lying the enamel to castings.

542 JOURNAL OF THE

Various types of screens are used, revolving screens probably
being the most common. In order to get the enamel powder
through these they must be jarred. Formerly, this was conmmoly
done by an iron hammer hitting on a ring of iron knobs encircling
the center of the screen. With the introduction of antimony
enamels, it was found that the contamination of the enamel by
specks of metal from these knobs and the hammer was very
noticeable. The use of rawhide hammers reduced the dirt some-
what and changed the specks from iron specks to carbon specks,
produced by particles of rawhide. Various expedients have been
employed to overcome this trouble with revolving screens, the
most practicable of which seems to be the placing of the beater
ring and hammer at one end of the screen frame, beyond the
screening area. This eliminates the dirt but cuts down the
capacity of the screen, since the jarring action is greatly reduced.

Another screen is of the reciprocating type. A flat screen is
hung from a frame by wire cables and made to move back and
forth by a short-stroke air motor such as is used for foundry
riddles. Such a screen is clean and well suited to plants of
moderate capacity. To keep down dust, it may be enclosed in a
cabinet, the motor being placed on the outside. Of course, the
air motor may be replaced by a belt-driven eccentric or any other
mechanical device that will give a reciprocating motion.

A sifter adapted from the flour milling industry has a combined
rotary and sliding movement. An oblong screen is set in a box
with a solid bottom. One end of this box runs on rollers, while
the other end, which is slightly higher, is supported by a ball
and socket joint on the end of a pin set eccentrically on a horizon-
tally running wheel. The sifter is operated by power applied to
this wheel by an adjustable friction drive. As the pulley revolves,
the upper end of the box takes a rotary motion while the lower
end runs back and forth on the rollers. The horizontal pulley,
and of course the upper end of the box, make about two hundred
and thirty revolutions a minute. The enamel passing through
the screen slides down the inclined bottom of the box and out.
through a chute at the lower end. The coarse material passes
out through a similar chute at the lower end of the screen. Since

AMERICAN CERAMIC SOCIETY. 543

11 wearing surfaces in this machine are below the floor of the box
1 which the enamel is sifted and caught, there is no contamina-
on of the enamel powder from that source. The capacity of this
:reen, as well as of the sliding screen previously described, may
e materially increased by attaching to the screen frame a com-
ressed air vibrator, such as is used for vibrating molds in
mndries.

THE ENAMELING PROCESS.
1 . Application of the Ground Coat.

After being sand-blasted and cleaned, the castings should be
ept in a dry, warm place until ground-coated, which should be
Dne within a few hours. The ground coat can be applied by
ainting, pouring, or spraying. In painting, a ground coat
>mposition of about the consistency of thick gravy is spread over
le surfaces to be enameled by means of flat painter's brushes — ■
»ur-inch brushes are used for tubs and two-inch for small ware,
he brush marks are likely to show through the enamel, especially

the piece is heated strongly during the firing of the ground coat
• enamel. The heat causes incipient blistering and swelling
' the heavier ridges left by the brush. For this reason, painting
' ground coats is not extensively practiced. Sometimes when
irtain parts of a piece are to have an especially thick coat, these
e painted before the piece is poured or sprayed.

In pouring, enamel of about the consistency of cream is poured
rer the casting. If the piece is small, it is turned in various
Dsitions and rocked backward and forward on two rods ex-
uding over a drip tub until the ground coat is smoothly on the
irface of the whole piece. The thicker beads, that accumulate
; the edges, are removed by running a finger along them. In
te case of tubs and other large pieces, the ground coat is poured
1 as uniformly as possible, then the piece is placed in various
)sitions and pounded with a rawhide mallet. The vibration
nds to cause the coat to run smooth. When done by a skilled
an, working with ground coat enamel of proper consistency,
>atings of very fair uniformity can be produced by pouring.

is a more rapid method than painting but not as rapid as spray-

si i JOURNAL OF THE

It is by spraying that the most uniform coats are applied, and
at the same time with the greatest speed. The spraying device!
are operated by compressed air, various types being sold for the
purpose. In addition, a number of kinds of home-made sprayers
are used in different plants. In general, it lias been found that
sprayers which depend on suction for their supply of coating give
finer-grained and more uniform sprays than those depending on
pressure. The sprays should be as fine grained as is consistent
with reasonable speed of operation. The ware is usually placed
on a turn-table while being sprayed. In some shops, small ware
is sprayed on a regular small-ware enameling table.

In spraying, only one-third to two-thirds of the material com-
ing from the spray nozzle lights on the ware. Therefore, the
spraying space should have a smooth wooden floor and smooth
wooden walls on three sides. If the castings are clean all over
before they go to the spraying cabinet and the workmen do not
carry in dirt on their feet, the ground coat on the floor and adher-
ing to the walls may be gathered up, sieved and used over again.
Dust from cement or brick walls and floors is liable to contaminate
the ground coat.

If an exhaust fan is used to carry away the vapors from the!
spray chamber, a large box or collection bin should be placed in
the pipe line. This will act like a dust collector and most of the'
material carried away will be deposited. It is best not to try;
to use this for ground coat alone but to mix it in small proportions
with fresh coat.

The following simple device for conserving labor and material i]
in the spraying of tubs is in use in several plants : The tub is
run on a truck into a three-sided wooden stall having a runway i
around the top of the sides, which extend a few inches above the
rim of the tub. The operator walks around the tub on this run-
way, spraying as he goes. The waste material is caught by tht
sides of the stall and may be used again.

After being ground-coated, the ware must be kept in a drji
warm place until enameled in order to avoid incipient rustinj 1
of the iron. This time should be brief, twenty-four hours at thi
most. Usually the pieces are stored near the enameling furnaces j

AMERICAN CERAMIC SOCIETY. 545

2. Enameling Room Equipment.

The enameling room is usually rectangular in shape with
lrnaces along one side. This room should be well lighted and
entilated and have a high roof, the lowest girders being at least
venty feet from the floor. A high roof not only makes a better
entilated room, but reduces the amount of enamel powder and
ust that lights on the girders and drops or is blown down from
me to time. Latticed monitors and in fact any kind of ventila-
>rs that can retain dust are also objectionable. The overhead
rders and underside of the roof should never be white washed,
; the coating peels off and spoils any hot enameled surface onto
hich it drops. When furnaces are built on both sides of the
>om it is likely to be dark and is always very hot.

The furnaces used for enameling cast iron are large rectangular
rens having an arched roof and a wide door at one end. If coal

used for fuel, the sides and arch are built double, forming a tight
uffle. If gas is used, the inner arch is sometimes omitted, thus
aking a semi-muffle. For tubs, the inside dimensions of the
uffles are about nine feet long, six feet wide, and five feet high
1 the crown of the arch. Small ware furnaces are made about
x feet square by four feet high, inside dimensions.

Various types of furnaces are used, most of the variations being
ade in the flue systems. Very elaborate flue systems have been
nbodied in furnace construction with the idea that the longer
ie length of travel of the flue gases in the furnaces the greater the
nount of heat given up. This, of course, is not necessarily true,
hese elongated flues are generally placed in the roof and sides

the furnace, which receive large amounts of heat by radiation
id conduction from the muffle and heating flues proper, and it
ippens often that the combustion gases leave the flues at a
mperature fully as high as that at which they entered. In
her words, heat cannot be withdrawn from gases by leading
iem through heated flues. In other cases, these flues are placed
rther away from the hot parts of the kiln and nearer the outside
irfaces. The gases then leave the flues at temperatures some-
hat lower than the entering temperatures. However, the heat
ven up simply goes to heat portions of the kiln exterior that are
irmally cool and to increase the radiation from these. The

546 JOURNAL OF THE

only way possible to extract heat in useful form from hot flue
gases is to use the hot gases for heating the air used for com-
bustion— by means of some form of regenerator or recuperator.

In general, it maybe said that the simple types of furnaces give
the greatest satisfaction. Complicated flue systems arc prone to
result in poor draft and frequent stoppage of flues with soot.
The more complicated the construction of a kiln, the greater is
the possibility of its going to pieces under the strains incident to
expansion and contraction of the furnace on heating and cooling.
Complicated furnaces are expensive to build, difficult to operate,
and expensive to keep in repair.

In Fig. 2 is given the plan for a tub furnace that has been
used with good results in several plants. Part of the flame
from each burner goes directly up along the wall and over the
top to the central collecting flue. The main portion passes under
the floor, up the opposite side, and over the top to the same
collecting flue. The distribution of the flame is controlled by the
dampers. By using a short flame, most of the heat is generated
under the bottom and in the lower ends of the side flues and the
products of combustion enter the collection flue at a comparativel)
low temperature. The material for this furnace can be had a'
reasonable prices since the construction calls for no special shapes
except the cup brick for the muffle, and these are used in largt
quantities by the enameling and other industries. As shown, th<
furnace is intended for the use of natural gas, producer-gas, o
oil. The burners are placed alternately on the sides of the furnace
When used for coal, fire boxes are added at one side, each fire bo
serving two flues, and all the firing is done from the one side. B;
removing the top arch of the muffle, this furnace can be readil
converted into an open arch, or, in other words, a semi-muffl
furnace.

In a few plants, double-chamber furnaces of this type are usee
the firing being all done on one side. One part of the produd
of combustion passes under the floor of both chambers and M
the far wall of the second. The other part passes entirely ov<!
the top of the first chamber, through the division wall near tl
floor level, and up the inner side of the second chamber. Tl
first chamber is always a tight muffle, but the second is usual

AMERICAN CERAMIC SOCIETY.

547

548 JOURNAL OF THE

operated as a semi-muffle furnace. The final melting of the
ground coat and all baking of enamels is conducted in the first
chamber, which is kept at ordinary enamejing furnace tem
peratures. The second chamber, which usually is at a dull red
heat, is employed for preheating the tubs, covered with ground
coat, to dull redness.

From a double furnace, more pieces can be turned out in a
given time with less expenditure of fuel per piece than from a single
furnace. However, since the double furnaces save nothing in
labor, this rapid rate of working is very severe on the workmen
and often causes dissatisfaction.

Another type of furnace which has come into use in a few
plants burning gas or oil has no fire boxes. Flames are turned
directly into the furnace chamber so as to heat it to a temperature
suitable for enameling. The supply of fuel is then shut off and
the tub is enameled simply by the radiant heat of the hot furnace
walls. As soon as the tub is removed from the furnace, the flames
are turned on again. This type of furnace is cheaply constructed
and very economical in the use of fuel.

In Europe, where fuel is very expensive, a number of types of
enameling furnaces equipped with recuperators have been built.
A few such furnaces have been constructed in this country, after
European designs. In Fig. 31 is shown the plans of a European
furnace with a recuperator built integral with it. It will be I
noticed that the large fire box is, when covered with a very thick
bed of fuel, really a gas-producer. This feature is one of the chiel
advantages of such furnaces. The air for combustion of tht]
gases produced travels around the tubes of the recuperator and isj
preheated by the out-going combustion gases. According tel
Damour,2 recuperation of secondary air for furnaces with a work j
ing temperature of 1800° F. will give a fuel economy of 30 pe;|
cent. These furnaces give good satisfaction, but it is a questioij
how much of this efficiency is due to the recuperative feature an(|
how much to the gas-producer and heavy, heat-retaining con J
struction in general. The first cost of these furnaces is quit!

1 E. A. Schott, Stahl u. Eisen, 30, 1556.

2 Emilio Damour, "Industrial Furnaces," p. 127. (Published by Engl
neering and Mining Journal.)

AMERICAN CERAMIC SOCIETY.

549

550 JOURNAL OF THE

high usually about double the eost of a simple direct-fired
furnace.

The fuel consumption of any of the foregoing furnaces can be
materially decreased by covering them with a layer of some of the
heat insulating materials now on the market. In regard to irons
and braces, it may be said that the more substantial these are the
less the cost for repairs to the furnace. In regard to the use of
open arches versus tight muffles — it is to be considered that where
open arches call for much less fuel than closed arches, their use is
more likely to give ware of poor finish and contaminated with dirt
from the roof of the furnace.

The doors to the furnace may swing sidewise, rise up, or drop
down. Swinging doors, whether single or double, suck out hot air
when opened and push in cold air when closed. Lift doors are liable
to cause dirt to be dropped onto a piece of ware being taken out of,
or put into, the furnace. Dropping doors are the cleanest and do
not cause air currents. The only disadvantage is that it is not
possible for the enameler to open them a little way in order to
observe a piece of ware in the furnace. Of course, opening a
door to look at ware is merely a matter of habit and convenience,
for the piece can always be inspected through a peep-hole left in
the door. Rising doors are most commonly used. In most plants,
compressed air is used for operating tub-furnace doors — a simple
hoist with a wire cable running over sheaves being the common
apparatus. In some plants electricity is employed. Occasionally
power is used to open small-ware furnace doors, but the mon
common practice is to have them opened by hand. Power >
operated rising and dropping doors are counter poised so tha
they are self-opening, the power being used simply for closinj
and keeping them closed. This greatly reduces the mechanica .
problem of operating the doors and at the same time gives assur -
ance that the furnace will be open for the extraction of ware il
anything goes wrong with the power.

Supporting the Ware. — The ware, while being enameled, i ;
placed on a rack or table, known as an enameling table. Thil
table is fitted with gears and levers which permit the top beinil
tilted at any angle to the perpendicular, or even turned upsic'l
down in some types, and revolved while in the various position j

AMERICAN CERAMIC SOCIETY. 551

Any good master-mechanic can design such a table and conse-
quently there are many varieties being used, most of them being
indigenous to the shops in which they are found. In general,
the tables are tilted on the main horizontal axle of the frame and
revolved around a pin extending through this axle. In tables not
intended to have the top turned completely over, the center of
rotation is placed between the two bearings of the main axle.
In tables whose tops are to be completely reversed, the center of
rotation is placed on an end of the main axle entending beyond
both bearings. Similar tables are used for enameling small
ware but in most cases these are operated by hand.

Formerly a hand crank was used for rotating enameling table
tops, but in the more modern plants the tops are rotated by
power. One type of table uses a reciprocating air motor, which
produces a rotary movement by means of a series of cams articula-
ting with gears. Such tables are jerky in their action. A much
more satisfactory method is the gearing of a rotary air-motor or
small electric motor to the mechanism ordinarily used for rotating
the table top by hand. The top should rotate at a speed of ten
to twenty revolutions per minute. With an air motor the speed
can be largely regulated simply by the amount of opening of the
foot- valve used. The tilting is still commonly done by hand,
but in a few shops this is done by the use of a small air hoist and
suitable sheaves a nd pulleys.

For supportin g the ware in the furnaces while being heated,
various sorts of racks, called "bucks," are used. These are
generally made of heavy sections of cast iron. Sometimes old
railroad rails are used for bucks for tubs and, less frequently,
blocks of fire clay. The latter do not warp nor deteriorate by
oxidation like metal but are difficult to shift about into proper
positions in hot furnaces. The best form of buck for tall cylindri-
cal pieces, such as lavatory pedestals and drinking fountains, is a
heavy base with an extension curved so that a long finger points
horizontally toward the door of the furnace. The pieces are
fired on this finger and, being in the horizontal position, ar e
heated much more uniformly than when fired in an uprigh t
position.

The ware is put into and taken out of the furnaces by long

552 JOURNAL OF THE

tWO-pronged forks, which are suspended cither from a crane or
from a trolley running on an I-beam. .Sometimes dirt drops oil
the I-beam into ware, but the trolley System is steadier and
easier to handle than the cranes. In a few plants the forks are
pivoted on a truck running on tracks extending out from the
front of the furnaces. This is a clean system, but does not permit
of ready adjustment of the fulcrum point of the fork to ware of
different heights.

Application of the Enamel. — Enamel powder is kept more
free from dust and is more readily obtained by the operators
working at high speed while enameling, if it is kept in a covered
receptacle from which it can be made to feed semi-automatically
in fixed quantities into the enameling sieves or dredges. Such
contrivances are generally known in the shops as "powder dum-
mies." One form of dummy consists of a large funnel with a
drum at the bottom divided into six wedge-shaped compart-
ments. The drum rotates on its horizontal axis and, except
at the bottom and top, is entirely encased by close fitting walls.
Whenever the contrivance is not being operated, one of these
compartments forms the lower end of the large funnel. When
he needs enamel powder, the operator pushes with his sieve on a
rod reaching below the drum, the movement of this rod operating
a pawl which causes the drum to rotate one-sixth of its circum-
ference. In this way one of the compartments is emptied into
the sieve.

A much more simple form of dummy consists of a large circular
funnel with a tubular extension on the bottom. This extension
is closed by a swinging slide with a weighted projection at the
bottom. To operate this dummy, the enameler pushes with his
dredge on the weighted projection. The slide is thus pushed
back and the sieve brought into position under the tube, the
enamel in the tube immediately dropping into the dredge. When
the pressure is removed, the weight on the projection causes the
slide to close. The enamel in the funnel proper slowly drops down
to fill the tube again. This simple contrivance is much less likely
to get out of order and costs less for repairs than any of the more
complicated dummies.

AMERICAN CERAMIC SOCIETY. 553

For protecting tubs from drafts during cooling, covered stalls,
died "dog houses," just large enough to hold one tub on a truck,
■e provided. The front of these stalls may be open or may be
rovided with loose fitting doors. Tight doors do not allow a
lb to cool fast enough and thus favor the growth of minute
ystals on the surface of the ware, which cause dull finish. These
alls may be built of any solid material. Sheet iron is not
litable, especially for the roof. The light sections of metal
uckle with heat from a hot tub and throw dirt onto the still
>ft enamel. It is not customary to use cooling chambers for small
are.

The dredges used for sifting the enamel onto the ware consist
I circular sieves attached to handles about five feet long. In
•der to have the enamel pass through the screen in an even
ream, the sieves must be vibrated. In a few shops this is
ill done by rapping the wooden handle of a dredge with a beater,
he beater consists of a small oval hoop of about 3/8" round iron
iving a wooden handle fastened to one end. The oval is placed
-ound the handle of the dredge and worked rapidly up and
Dwn by the right hand of the enameler while the dredge is held
I his left hand, and, in the case of large sieves, partially sup-
3rted by a counterpoise. The more common method of vibra-
ng dredges is by means of a small air or electric vibrator at-
iched to the end of a gas-pipe handle on the dredge. The
)mmon form of air vibrator is a small chipping hammer with a
lunt point substituted for the chisel. This may be a long rod
ith its tapping end beating against a plate attached to the
all of the sieve, or it may be very short and beat against a
ardened disk held in a case at the end of the dredge handle,
ince the whole dredge, including the handle, must be vibrated
1 either case, the point of application of the blows is obviously
nmaterial. Electrically operated vibrators of several types
ave been used, but they have not proven as satisfactory as the
ir hammers. The introduction of mechanical vibrators for
namel dredges has been a great boon to the enameling industry,
'hey enable the workmen to produce more and better ware with
luch less effort. Formerly, paralysis of the hand used for beat-
lg was not uncommon among enamelers. For hand vibrated

554 JOURNAL OF THE

dredges, 30 to 40 mesh wire gauze is used, while lor power-
vibrated, 40 to 60-mesh is employed.

3. Enameling a Bath Tub.

Having previously been coated with ground coat, which must
be perfectly dry, the tub is caught up along the under side of the
rim by one of the large forks, operated by the two enamelers,
and placed in the furnace. In from seven to ten minutes, the iron
is a bright, cherry-red, and the ground coat has melted evenly.
The piece is quickly withdrawn and placed on the revolving table;
One man sifts enamel on the rim through a small sieve while the
other man dredges the sides and bottom. The heat of the iron
fuses the enamel enough to make it stick to the surface. In
about three minutes the piece is thoroughly covered, and is im-
mediately put back into the furnace. In a few minutes the enamel
has melted smooth, and the tub is drawn and given another coat.
This runs smooth in a few minutes, and the piece is again drawn.
Any blisters appearing are patched by being perforated by a steel
point and covered with a little mound of enamel powder. The
tub is given a third light coat and allowed to bake. It is then
drawn, any enamel that has run over the edge of the casting is
trimmed off with a large knife, and the piece is allowed to cool
in a "dog house."

The proper enameling of a tub is a matter of considerable skill
and judgment. Each dredge must be evenly distributed over]
the piece. If it is too thin in places, blue spots show on the
finished ware; if too thick, crazing or chipping off of the enamel
may occur. One fault especially difficult to avoid is the tendency
of the enamel to run and collect in rolls or beads at certain points.
This spoils the appearance of the piece and often results in crazing
when the tub cools.

The temperature at which practically all tub enamels mature
is very close to 10000 C (18000 F.). It may not be much higher
for between 10500 and noo°C (1900 and 20000 F.), cast iroij
burns "dead," becoming exceedingly weak and brittle. It can J
not be much lower on account of the increased cost. An ename
that would mature much earlier would have to be considerably i
higher in the more expensive fluxes. Another important con

AMERICAN CERAMIC SOCIETY. 555

sideration is that an enamel which would mature much earlier
would be a softer glass, more readily soluble, and more liable to
deteriorate under the ordinary conditions of use. An enamel
for cast iron that will meet all the requirements of the service for
which it is intended can be made at 18000 F.

Though the enamel matures at 18000 F, the furnace is fired to a
considerably higher temperature. The temperature is kept down
by the process of enameling. Every thirty minutes, or there-
abouts, a cold tub weighing from 200 to 300 pounds is put in,
heated to 18000, drawn out three times, allowed to cool con-
siderably and heated up again after each cooling. This process
keeps the furnace and the men busy, insuring a maximum output
of ware, but it calls for skill and watchfulness on the part of the
enameler. Each coat of enamel is baked very quickly; it is
melted down in from two to five minutes in a rapidly rising heat.
Exposure for a minute too long in the furnace during any one of
the four heating periods may ruin the piece. The defect may be
burning up of the slush coat, sulphur bubbles, running off of the
enamel, etc. If the tub is not heated sufficiently, the result will
be pin holes in the coldest parts, and' a rough, uneven coat, of poor
luster.

After the tub is cold, it is carefully inspected as to defects in
application and fit of the enamel, color, luster, etc. Minute
crazes are readily detected by rubbing the surface with a rag
filled with black grease. The tub is then painted on the out-
side, supplied with feet and fittings, and taken to the ware room.

A PYCNOMETER OPERATED AS A VOLUMETER.1

Bv H. G. Schurbcht, Columbus, Ohio.

( )wing to the fact that in studying the drying and burning
behavior of clays, the volume shrinkage can be determined more
accurately than the linear shrinkage,2 a satisfactory volumeter
for measuring this shrinkage is an important piece of apparatus
in a ceramic laboratory. Heretofore, the Seger volumeter,'
although unsatisfactory, has been used almost universally. Some
of the objections to the Seger volumeter are enumerated as fol-
lows:

i. It is too time consuming. In order to obtain accurate
results with a Seger volumeter, it is necessary to allow the liquid
in the burette to drain at least 3 minutes before making a reading.
The total time required for a single determination of volume
varies from 5 to 10 minutes.

2. Particles of grit or clay may become lodged in the large
stopper joint — thus changing the volume of the bottle when
closed.

3. It is quite fragile and easily broken.

4. Errors are often introduced through the entrapping of air
bubbles.

Pycnometer Volumeter. — In order to overcome some of these !
objections a volumeter (Fig. 1) operating on the principle of the
pycnometer was used in the Bureau of Mines Laboratories with
excellent results. The bottle is made of glass — metal being con-
sidered too heavy as it would decrease the sensitiveness of the
balance and thereby decrease the accuracy of the determination.

1 By permission of the Director, U. S. Bureau of Mines.

2 J. Bailey, "Variation in Linear Shrinkage of Clay Test Pieces," Trans.
Am. Ceram. Soc, 18, 557-563 (1916).

3 For details regarding operation of the Seger volumeter see H. Ries,
"Clays, their Occurrence, Properties and Uses," Robert Drummond & Co.,
N. Y., pp. 160-62.

AMERICAN CERAMIC vSOCIETV

557'

bV"T"^ Joint
HL Ground
F ro fit

VOLUMETER

Glass ; opera ted
pofcnonneter

as a

Section Through A A

Fig. i.

558 AMERICAN CERAMIC SOCIETY.

Operation, 'flu- bottle is filled with liquid (kerosene or water]

and, as the Stopper is inserted, sufficient liquid is forced into the
tube in the stopper so as to lill it completely. The excess liquid
is carefully wiped off and the bottle is weighed. The briquette
to be measured is then inserted into the bottle and the operation
is repeated. The calculation involved is as follows:
_ W — W, + B
S

V = Volume

W = Weight of bottle plus liquid

Wi = Weight of bottle plus liquid plus briquette

B = Weight of saturated briquette

S = vSpeeific gravity of liquid

It is necessary to cheek the weight of the bottle plus the liquid
at intervals as minute particles of clay, dropping from the bri-
quettes, may increase this weight.

A mark is scratched on one side of the stopper and bottle so
that the stopper may be placed each time in the same position
with reference to the bottle — any error which might be caused
by inserting the stopper in different positions being thereby
eliminated.

In comparing this type of volumeter with the Seger volumeter
it might be said:

i. That it is much more rapid, in that a determination can be;
made in from i to 2 minutes as compared with the 5 to 10 minutes
required when using the Seger volumeter.

2. It is very accurate. Volumes can be deterimned to within
0.01 to 0.05 cc. — the accuracy depending on the size of thei
briquette and the sensitiveness of the balance used.

3. The error due to the lodging of grit or clay in the joint
of the stopper and to forcing the stopper further down into th( ]
bottle in one case than in another would be less in this apparatu;
than with the Seger volumeter — as the stopper of the former i:l
smaller and the displacement would be smaller in proportion.

4. There is less chance of error from air bubbles.

5. It is not as easily broken as the Seger volumeter.

6. It is less expensive.

COMMUNICATED DISCUSSION

Bv F. Gelstharp,
OF THE PAPER

THE EFFECT OF CERTAIN IMPURITIES IN CAUSING MILKINESS
IN OPTICAL GLASS."

By C. N. Fenner and J. B. Ferguson.1

The Pittsburgh Plate Glass Company has been very suc-
essful in producing the several types of optical glass. Al-
hough in our early work we occasionally encountered milky or
palescent glass, either in the pot or after molding, we have not
bserved the phenomenon for a long while. The only glasses in
/hich this opalescence was produced were the borosilicate crown
refractive index 1.522) and dense flint (refractive index 1.6 16).
mpure potassium carbonate was not the cause of the
palescence because the pure nitrate was used in place of car-
lonate. I do not feel that the authors of this paper have posi-
ively proven that the presence of sulphates in certain optical
;lasses will produce opacity or milkiness.

This is a very interesting subject to glass manufacturers gen-
rally as opalescence has been the cause of trouble in other branches
if the glass industry. When ordinary soda-lime glasses turn
nilky or slightly opalescent, the milkiness may invariably be
liminated by a slight increase in the alkali content of the glass.

From the small amount of research work I have done on simple
passes, I believe that I am correct in stating that, having a
upersaturated solution of silica, the tendency is for this to
>recipitate out, either on cooling slowly, or, if the cooling is too
apid, it will separate on reheating. I therefore propose a
lifferent view for the solution of this problem.

The optical glass referred to by the authors and those glasses

vhich I have mentioned contain a very high molecular content

)f silica and therefore approach supersaturated solutions. If

or some reason the alkali content is lower than intended, milk-

1 J. Am. Ceram. Soc, I, 468-476 (1918).

560 I1 >URNAL OF THE

iness may easily result. I would offer the following explanations

for this:

First: In the case of Russian potash, if a batch were made
up containing 68 parts by weight of potash (calculated to anhy-
drous KoCO.i assuming that the KC1 and K2S()4 present will
supply their quota of K2O) to 300 parts of sand, it is very probable
that there would be a deficiency in alkali sufficient to produce
a supersaturated solution. In case melts from this batch did
not opalesce, probably a higher temperature prevailed and the
K2S()| dissociated or reacted with the silica in solution, yielding
its quota of K20, this reaction being more or less complete. Some
of the KC1 may have reacted and supplied some K20, part may
have remained in solution as KC1, and the remainder may have
been volatilized.

Second: Bearing in mind the first probability, there was
evidently considerable variation in the composition of the Russian
potash which might account for some of the melts showing opacity.

Third: We have found that a considerable amount of alkali
may be lost by volatilization and that the amount varies with
furnace conditions. This might also account for the erratic
results secured when the proportions are so near a supersaturated
solution.

A case is cited by the authors in which melts were made in two
pots in the same furnace, one producing milky and the other
clear glass. The respective content of S03 is stated to have
been 0.10 per cent and 0.05 per cent, respectively — the CI content
being the same in each melt. It is remarkable if such a slight
difference could possibly account for the conditions of opalescence
— I feel that some other reason must be found to adequately
explain it.

I feel convinced that, in the case of the borosilicate crown
and dense flint glasses to which I referred, the milkiness was
due to a deficiency of alkali — due to volatilization caused by
improper melting conditions. I suspected that the arsenic was
responsible for the cloudiness but subsequent melts in which a
larger amount was used in the batch did not develop it. In one
of our early experiments an excessive amount of arsenic \va>
introduced by mistake and a dense white glass resulted. This

AMERICAN CERAMIC SOCIETY. 561

was also probably due to the formation of a supersaturated solu-
tion of silica on account of the fixation of a portion of the alkali
by arsenic.

I have never found that the presence of SO3, CI or F in
soda-lime glass (in amounts up to 0.6 per cent of each) caused
opalescence to the slightest extent, so long as the ratio of alkali
to silica was below the limit at which opalescence occurs when
SO3, CI or F are absent.

In the case of opal glasses, such as the so-called Carrara glass,
fluorine is generally used in some form, but its efficiency in pro-
ducing opacity seems to depend on both the form in which it
is introduced and the ratio of alkali and other bases to the silica.
It has never to my knowledge been positively proven whether
the opacity in this case is due to separated silica or to some fluorine
compound.

I do not wish to convey the impression that the authors are
incorrect in concluding that the presence of SOs and CI in some
optical glasses will cause milkiness, but simply to suggest that
before this can be accepted, further experimentation is necessary.
However, I believe it is very probable that the presence of salts
in solution in glasses which contain a high proportion of silica
(approaching a supersaturated solution) will tend to cause such
glasses to more readily turn milky on slow heating or cooling.

Pittsburgh Plate Glass Company,
Creighton, Pa.

Reply to F. Gelstharp's Discussion,

By C. N. Fenner and J. B. Ferguson.

We are very glad that Mr. Gelstharp has brought up for con-
sideration the possibility that the opalescence or milkiness which
we have described was caused by a deficiency of alkali. This
is a reasonable hypothesis and deserves careful consideration,
but the evidence does not seem to us to support it. We wish,
however, to emphasize again the statement which we made in
the original article, that we do not intend the explanation there
offered to be understood to apply to all cases of milkiness or
devitrification in glasses. It is only reasonable to suppose that
a deficiency of alkali in a glass would be a contributor}- cause,
and under some circumstances might be the most important

JOURNAL OF THE

canst' leading to devitrification. Nevertheless, we are strongly
of the opinion that many glasses arc somewhat supersaturated
with silica and still do not ordinarily turn milky or devitrify,
but that other factors may come in which exert controlling in-
fluences and lead to this result.

In what follows we wish to restrict ourselves to the phenomena
which we discussed, and to give reasons why Mr. Gelstharp's
interpretation does not seem to apply.

i. A suggestion is made as to the manner in which the use
of a potash carrying S()3 and CI might lead to a deficiency in
alkali. This is based upon a perhaps natural assumption as
to the manner in which we allowed for the presence of these
ingredients in our calculations, but as a matter of fact the pro-
cedure that we followed had a tendency to increase rather than
decrease the percentage of alkali.

As we have mentioned, the maximum S03 content of Armour
potash which it was considered judicious to tolerate for the light
flint glass was 0.30 per cent, and with even this amount there
was found to be a tendency toward the production of milkiness.
That this was not due to a deficiency in alkali is shown by a
consideration of the method used in adjusting our calculations.
This was as follows: Since there is no reason to suppose that
the K20 content of K2S04 in a lead-alkali-silica mixture is more
volatile than the K20 of K2C03, its equivalent in terms of K2C03
was calculated and used as such if the K2S04 content was large,
but if it was only 0.20-0.30 per cent (as was generally the case)
it was believed that no essential difference in the glass would
be made by reckoning the K2SC>4 as K2C03 without conversion.
The effect of this, as regards the composition of the glass, was
almost imperceptible, as the following calculation shows:

Suppose the raw potash contains 80 per cent anhydrous K2COa
and 1 per cent K2S04 (equivalent to 0.46 per cent S03).

If all the K2S04 were called K2C03, this would make 81 per
cent K2C03 in the raw potash, and to get 68 kg. of anhydrous
K2C03 in the batch 83.95 kg. °f raw potash would be used. This
would be supposed to contain 46.376 kg. of K20. Taking ac-
count of the K2S04, however, the 83.95 kg- °f raw potash would
actually contain 45.803 kg. of K20 from carbonate and 0.454

AMERICAN CERAMIC SOCIETY. 563

<:g. of K20 from sulphate, or a total of 46.257 kg. of K2() instead
)f 46.376 kg.

A batch of the kind under consideration would yield about
590 kg. of glass. The percentage of K20 in the glass in the two

>ases would be

(a) 46.257 -j- 500 = 7.84 per cent.
(,b) 46.376 -r 590 = 7.86 per cent.

rhat is, the error (deficiency) arising from considering the K2SOt
is K2CO3 would amount to 0.02 per cent K2O in the final glass,
m insignificant amount.

KC1 was treated differently. Data on what happens to KC1
n a batch of this kind are very meager, but as we know that
iCCl is highly volatile, and as the amount was small, it seemed
;afest to consider that all of it escaped and not consider it a
iource of K20 at all. Possibly some objection might be raised
o this procedure but the reason for it would be the converse of
hat which Mr. Gelstharp has suggested. If any of the KC1
lid become converted into a K2( ) compound and remain in the
nelt, the effect was to raise the K20 content.

2. Mr. C.elstharp speaks of the probability that the glasses
ipproach a supersaturated solution, and there seems to be an
mplication here that, as regards composition, they are in a critical
egion and that a little addition of silica will throw them over
nto a condition of supersaturation, causing a precipitate of
iilica to appear. We hesitate to believe that such is the state of
tffairs. We are rather inclined to think that all glasses of the
ype under discussion are somewhat supersaturated at 950
[ooo° C and that a few tenths per cent increase or decrease
n silica has no effect. A certain set of laboratory experiments
vhich we have lately carried out gives evidence on this and other
)oints.

A glass which became milky on reheating was used as the basis
or the experiments.

Experiment 1 . Several chunks of glass were placed in a platinum

rucible and heated for about an hour at 950-1 000 ° C. The

' Necessarily a definite temperature must be expressed; otherwise super-

-ituration has no meaning. We have taken a temperature at which latent

lilkiness is developed on reheating.

564 JOURNAL OF THE

glass was cooled and broken out. It appeared very milky;
even small pieces, 2 3 mm. in size, showed very perceptible
opalescence.

Experiment 2. Another sample of the same glass was ground

to pass 80 mesh and heated as follows:

1 .50 p.m. 1410 C 2.15 p.m. 1490 (

3 45 P.M. 1 1 ■ tin P.M. 140 I (

The glass, cooled quickly, was perfectly clear. To determine
whether milkiness could be developed it was reheated as follows:

liq.i A.M. 9l6° C II.22 A.M. 1020° C

I2.08 P.M. 996 ° G I Jo I'M. 1008 ° C

It was then removed and cooled. The conditions were the opti-
mum for developing milkiness, but none appeared. On the
surface there was a thin film of white, de vitrified material, but
the bulk of the glass appeared perfectly clear.

Experiment 3. Another sample of the same glass as No. 1
was ground to pass 80 mesh and mixed thoroughly with 2 per
cent of finely ground quartz. The heating was from 14000 C,
at 2.25 p.m., to 15000 C at 4.18 p.m. It was cooled, and then
reheated at 1018-10420 C for about three hours. After this
treatment the surface devitrification was a little greater than in
Experiment 2, but the material underneath the surface appealed
to be wholly free from milkiness or other form of devitrification.

From these experiments we infer that the original sample of
glass contained some ingredient which assisted the excess silica
to crystallize out and thereby milkiness was produced, but that
when this glass was heated to approximately 15000 C this in-
gredient was volatilized, and the glass, although still super-
saturated with silica at 950-1 000 ° C, thereby lost its power to
free itself of excess silica. This result can hardly be attributed
to a change in alkali produced by the high temperature, for any
change of this sort would certainly be in the direction of a decrease
rather than an increase in alkali.

AMERICAN CERAMIC SOCIETY. 565

3. Although laboratory experiments of this kind tend to sup
ort our conclusions, we do not attach nearly so much importance
3 them as to the large-scale experiments which the commercial
perations represent. These showed the striking fact that when
roper measures were taken to maintain the furnace temperatures
t the desired level, the appearance of pots of milky glass ceased
nmediatelv, and for months following, trouble of this kind was
ractically negligible.

It is difficult to understand how this procedure can be supposed
) cause more of the somewhat volatile alkalies to be retained
1 the melts, but it is easy to see that it would tend to drive out
le last traces of any such materials as SO3 or CI.

Before making the changes in methods of temperature-con-
'ol which we have described, Mr. Victor Martin, then chief
lass -maker at the Bauseh and Lomb plant and a man of wide
tperience in practical glass-making, had expressed the opinion
lat the trouble arising from milkiness could be obviated if

were possible to operate the furnaces at a higher temperature,
ater experience showed that this was essentially true but that

was only necessary to keep the temperature at a point very
ttle above that which we had believed to be suitable for ordi
By conditions but which at times was not attained because of
■peptive thermoelement indications.

Mr. Celstharp refers to certain observations of his on the
feet of arsenic in producing milkiness. He believes that the
'senic fixes a certain amount of alkali and thereby increases the
nount of uncombined silica. Even if this be true, the amount

alkali which can reasonably be supposed to combine with the
nail amount of arsenic which is sufficient to produce milkiness is so
nail as to be practically negligible. To us it seems at least
; reasonable that the action of arsenic is similar to that which
e have assigned to S03 and CI.

A criticism which seems to us applicable to Mr. Gelstharp's
)int of view is that in drawing his conclusions he does not make
ifficient distinction between the amount of S03, CI and As which

put into the melt and that which stays in. All of these ma-
rials, in whatever form they are added, tend to be volatilized

high temperatures, and the statement that in certain cases

566 AMERICAN CERAMIC SOCIETY.

a relatively large amount of one or the other went into the milt
without producing milky glass has, of itself, little bearing upon
the question. We need to have in addition fairly complete in
formation on the subsequent heat treatment or an analysis of
the glass.

Although we agree with Mr. (k-lstharp that final proof of our
contention has not been obtained, nevertheless the evidence
in that direction appears to us very strong, and no other ex-
planation appears to meet the requirements.

Geophysical Laboratory.

Carnegie Institution of Washington,

Washington, D. C.

HEAT BALANCE ON A PRODUCER-GAS FIRED
CHAMBER KILN.

By R. K. Hursh, Urbana, 111.

The data presented here are based on tests made at the plant
! the Baker Clay Company at Grand Ledge, Michigan, in the
immer of 191 7. The equipment tested consisted of a kiln of
xteen chambers, each having a capacity of fifty to fifty-five
lousand standard sized brick, and three six-foot, water-sealed
is-producers of the pressure type.

The producers were connected by a header with the main
as-duct leading to the kiln. Each producer was equipped with

blower of the injector type using steam at pressures ranging
■om fifteen to forty pounds (gauge). A Virginia lump coal in sizes
p to four inches was fired in charges of one hundred and fifty
5 two hundred pounds to each producer at hourly intervals.

Air for combustion in the kiln was drawn through the open
oors and crown holes several chambers back of the one on fire
nd the combustion gases were passed through three heating
hambers before delivery to the stack flue. No use was made
f the usual water-smoking flues on the top of the kiln. The
verage firing period was forty to forty-four hours, the finishing
2mpcrature being 1075 ° C (19700 F.) in the top of the chamber.

Tests.

Coal was weighed separately for each producer and an average
ample was reserved for analysis. Ashes were removed from the
its every other day and a sample was taken for analysis. Both
nalyses were made by J. M. Lindgren at the Chemistry Dept.
f the University of Illinois.

The steam pressure at the blowers of the producers were taken
t hourly intervals and the diameter of the orifice in the blower
fas measured. From the average pressure and the area of the
fifice, the steam consumption was calculated by means of the

568 JOURNAL OF THE

Grashof1 formula. The use of a steam meter and a throttling
calorimeter to determine the quality of the steam would havl
been desirable, hut these were not available at the time. Since
the pressure of the steam was reduced through a valve before
reaching the gauge and the needle-valve at the orifice, it was as-
sumed that the steam was saturated at the pressures read. Un-
der the conditions this would not be greatly in error.

vSamples of the hot producer-gas were collected through a
glass lined sampling tube inserted in the gas-main a short dis
tance from the header. The tube projected into the main to a
point estimated to represent the mean velocity of gas flow. The
gas was collected in glass containers over water saturated with
the gas. To obtain an average sample of the gas, the sampling
period extended over a time of one hour, or from one firing of
the producers to the next. Analyses were made by means of a
portable Burrell apparatus, C02, 02, unsaturated hydrocarbons
(considered as C2H4), CO, H2, CH4 and N2 (by difference) being
determined. Owing to the difficulties of accurate sampling and
analysis, no attempt was made to determine the soot and tar.
The losses due to separation of these in their passage through
the ducts to the firing chamber were therefore accounted for in
the item of general losses.

The temperature of the producer-gas was measured at the
point of sampling by means of a base metal thermocouple encased
in an iron tube.

The temperature and humidity of the atmospheric air were
determined by means of wet and dry bulb hygrometer readings.

Temperatures in the tops of the kiln chambers were measured
by means of a platinum platinum-rhodium thermocouple inserted
through the crown holes. These readings did not represent the
average temperatures of the ware in the chambers but gave a measure
of the rise in temperature during the test except in the last cham-
ber through which the combustion gases passed. There the
average rise in temperature of the ware was necessarily esti-
mated from the data obtained.

Flue-gas samples were taken from the main draft flue near the
point of delivery from the kiln. An iron sampling tube was in-
1 Marks, "Mech. Engrs. Handbook," p. 354.

AMERICAN CERAMIC SOCIETY. 569

rted with the end opening at the point of approximately mean
s velocity. With the low temperatures of the flue-gases, it
is considered safe to use an iron tube for this purpose. The
mperature was measured at the same point with a silver-con-
mtan thermocouple.

An attempt was made to determine the average velocity of
e flue-gases in the main draft flue by means of a Pitot tube, but
e manometer was not sufficiently delicate to give satisfactory
adings. The volume of flue-gas was therefore calculated from
; analysis, from the analysis of the producer-gas, and from the
sight of coal fired.

The test covered the period of firing one chamber — forty-four
»urs from the time the fires were started in it until its finish
id the beginning of firing in the next chamber. Data on the
ing of several chambers indicated that this test might be con-
iered as fairly representative of the kiln operation.

CALCULATIONS.

Producer.

Volume of Producer Gas. — The weight of carbon consumed

:r kg. of coal fired was the carbon in the coal less that in the

, o. 1052 X 0.0472

hes, or 0.7437 — = 0.7375.

0.8028

The carbon per cu. m. of producer-gas was:
CO2 = 4.02 % = o . 0402
C>H4 = 0.6 % =0.012
CO = 22 .5 % = 0.225
CH4 = 1 .4 % = 0.014

o .2912 X = o . 156 kg. per cu. m. of gas

22 .4

O.7375

~ = 4.73 cu. m. producer-gas per 1 kg. of coal fired.

Volume of Dry Air Used. — The N2 content of the producer-
is furnished a measure of the amount of air used in the producer.

0.6348 X 473 = 3 005 cu. m. N-i in gas per 1 kg. of coal.

22 .4
0.0139 X = o .01 1 cu. m. N2 in gas from kg. of coal.

28

2 .994 = cu. m. N2 in gas from air used per i kg. of coal.

,S7<> JOURNAL OF THE

2 . W4

" = 3-79 cu. m. dry air used per kg. of coal (at X. T. I'.).
0.79 "■ '

3.79 X 1 .293 = 49 kgs. dry air used per kg. of coal.

The average temperature of the atmospheric air was 1 7 . 5 ° C
and the average moisture content was 0.00801 kg. per kg. of
dry air. The total moisture entering the producer with the
air was :

4.9 X 0.00801 = 0.039 kg. per kg. of coal

Steam Used in Producers. M = 0.0165 F p*97 (Grashof).

M = lbs. steam flowing per second.

F = area of orifice in sq. ins. = 0.0276.

p = absolute pressure of steam in lbs. per sq. in.

Multiplying M by 3600 gave the weight of steam per hour,
and dividing the total for the three producers by the average
weight of coal fired to the three per hour gave the weight of steam
in lbs. per lb. of coal, or in kgs. per kg. of coal.

Producer 1. M = 0.0165 X 0.0276 X 35. 5-97 X 3600 = 52.3

Producer 2. M = 0.0165 X 0.0276 X 34. 4'97 X 3600 = 50.7

Producer 3. M = 0.0165 X 0.0276 X 32. 297 X 3600 = 30.0

1330

133.0

—— — = 0.284 lb. steam per lb. coal or kgs. steam per kg. oi

468.5

coal.

The moisture entering the producers from the moisture and

combustion of the hydrogen in the coal was:

0.0258 + 0.039 X9 = 0.377 kg. per kg. of coal
The total moisture entering the producers from all sources was
0.039 + 0.284 + 0.377 = 0.700 kgs. per kg. of coal fired

Free Moisture in Producer-Gas. — The difference between th(
total moisture entering the producers and the moisture equivalent
to the free and combined hydrogen in the gas represents the fre<
moisture in the gas.

H2 =7 -5% = 0.075 cu. m. H20 vapor per cu. m. gas (at N. T. P.)
C2H4 = 0.6% = 0.012 cu. m. H20 vapor per cu. m. gas (at N. T. P.)
CH4 = 1 .4% = 0.028 cu. m. H20 vapor per cu. m. gas (at N. T. P.)
0.115

mi

AMERICAN CERAMIC SOCIETY. 571

0.1 15 X 4 .73 = 0.544 cu- m- H2O vapor equivalent to H2 in
as per kg. of coal, or 0.544 X 0.804 = 0.437 kg.

0.700 — 0.437 = 0.263 kg. free moisture in gas per kg. coal
red.

Calorific Value of Producer-Gas from Analysis.

C2H4 = 0.6% = 0.006 X 12915 = 77.5

CO = 22.5% = 0.225 X 3044 = 684.9

H2 = 7-5% = 0.075 X 2594 = 194.5

CH4 = 1.4%= 9.014 X 8558 = 119.8

1076.7 Cals. per cu. m. gas

1076.7 X 4.73 = 5095 Cals. per kg. of coal fired.

Sensible Heat of Producer-Gas. — The average temperature of
le producer-gas was 575 ° C. The average specific heats were
btained from the formulas of Langen.1

02 = 4.02% = 0.0402 X (0.39 4-0.00012 X 575) X 575 = 10.6

H4 = 1.4% =0.014 X (0.34 + 0.00036 X 575) X 575 = 4-4

2H4 = 0.6 % =0.006 X (0.42 -f 0.00049 X 575) X 575 = 2.4

H2, CO )

O N ( 93'98% = °-9398 X (0.31 + 0.000027 X 575) X 575 = 175-7

Sensible heat per cu. m. of dry gas = i93-i Cals.

193.2 X 4-73 = 914 Cals. per kg. of coal fired.
0.263 kg. H20 vapor, 0.263 X (0.44 + 0.000012 X 575) X
75 = 77 Cals.

The total sensible heat in the moist producer-gas per kg. of
>al is 914 + 77 = 991 Cals.
The total heat content of the hot producer-gas was:

5095 + 991 = 6086 Cals. per kg. of coal fired
Heat Input for Producers. — The heat content of the dry air
ipplied to the producers at the average temperature, 17.50 C,
as:

4.9 X 0.241 X 17.5 = 20.8 Cals. per kg. of coal fired
The heat content of the moisture in the atmospheric air at
ferage conditions was:

4.9 X 0.00801 X 604 = 23.7 Cals. per kg. of coal fired

1 Marks, "Mech. Engr. Handbook," p. 366.

57? J< >URNAL OF THE

The total heat in the atmospheric air supplied was 44.5 Cals
per kg. of eoal.

The heat content of tlje steam supplied to the blowers, as-
suming it to be saturated at the average pressure, was:
0.284 X 644 = 183 Cals. per kg. of coal fired.

The calorific value of the coal was 7108 and the total heat in-
put to the producers was 7108 + 183 + 44.5 = 7335 Cals. per
kg. of coal.

6086
The Hot Gas Efficiency was - - = 82.07 per cent.

7335

Heat Lost in Unburned Carbon in Ashes. — The amount of

unburned carbon in the ashes in terms of weight of coal was:

0.0472 X o. 1052

— ~ = 0.0062 and the heat value, 0.0062 X

0.8028

8100 = 50 Cals. per kg. of coal fired.

Undetermined Heat Losses. — The heat losses due to conduc-
tion, radiation, etc., were determined by difference.

7335 — (6086 + 50) = 1 199 Cals. per kg. of coal fired.

Kiln Data.

Volumes of Flue-Gas and Air Used in Kiln. — The C02 from
combustion of the producer-gas amounted to 2 . 9 per cent by
volume of the flue-gas. The volume of CO2 from the producer-
gas was :

C02 = 4.02% = 0.0402

C2H4 = 0.6 % = 0.012

CH4 = 1 .4 % = 0.014

CO = 22 .5 % = 0.225

0.2912 cu. m. C02 per cu. m. producer-gas

0.2912

= 10.04 cu. m. flue-gas per cu. m. producer-gas.

0.029

The N2 in the flue-gas amounted to 79.3 per cent or 0.793 X
10.04 = 7 962 cu. m. per cu. m. producer-gas. Of this amount
0.6348 cu. m. was from the producer-gas and 7.962 — 0.6348 =
7.33 cu. m. was from the air supplied in the kiln. The volume
of air furnished was:

AMERICAN CERAMIC SOCIETY. 573

7.33

- =9.28 cu. m. (at N. T. P.) per cu. m. of producer-gas.
0.79

The total volume of flue-gas during the period of the test was

9371 X 473 X 10.04 = 445>ooo cu. m. (at N. T. P.)

The total volume of dry air supplied to the kiln was

9371 X 473 X 9.28 = 411,300 cu. m. (at N. T. P.).

The weight of moisture introduced with the air supplied to

he kiln was 411,300 X 0.00801 X 1293 = 4260 kgs.

Sensible Heat of Flue-Gas. — The sensible heat of the flue-gas,

alculated by the use of the Langen formulas for average specific

leat, was:

44

C02 (0.029 X 445,000) X 0.21 X 170 X = 1,140,000 Cals.

22 .4

32

02 (0.178 X 445,000) X 0.217 X 170 X = 4,175,000 Cals.

22 .4

28

N2 (0.793 X 445,000) X 0.247 X 170 X = 18,530,000 Cals.

22.4

Sensible heat of dry flue-gas = 23,845,000 Cals.

The total moisture in the flue-gas from the air used in the kiln
nd from combustion of the producer-gas was:

4260 + 9371 X 0.700 = 10,820 kgs. Its heat content at
he average temperature of the flue-gas was 10,820 X 0.46 X
70 = 845,000 Cals.

The total heat in the flue-gas was therefore 24,690,000 Cals.

Heat Used for Burning the Ware.— During the test the tem-
erature was raised in four chambers of the kiln. The heat
sed in each chamber for raising the temperature of the clay
ras calculated separately. In one chamber the chemical water
1 the clay was expelled — requiring expenditure of heat — and
3me carbon contained in the clay was consumed. The heat from
tie latter was added to the other items of heat input for the kiln.
1 the last chamber some mechanically held moisture was ex-
elled.

Since the temperature of the flue-gases leaving the kiln was
Dnsiderably lower than that at which either the hygroscopic
r chemical water in the clay was expelled, the heat content of

574 JOURNAL 0* THE

the steam from both sources was represented by the heal in super-j

heated steam at the temperature- of the flue-gases and at the ex-
istent vapor pressure. This value was obtained from Good-]
enough's steam tables.

In the case of chemically combined water, the heat required
at dissociation would be that required for the endothermic re-
action plus the latent heat of vaporization of steam at the dis
sociation temperature. In the dehydration of clay, the tempera
ture of dissociation is above the critical temperature of water
and the latent heat of vaporization would hence be zero. How-
ever, it must be considered that, in heating up the clay to the dis-
sociation temperature, the moisture in the clay has taken up
an amount of heat equivalent to the heat content of steam at
the temperature and pressure existent at the point of dissocia-
tion. Hence, the use of formulas for the latent heat of vaporiza-
tion of steam at dehydration temperatures of clays is not strictly
correct.

Various values are given for the heat of dissociation of clay.
Bleininger1 used an assumed value of 200 Cals. per kg. of water
expelled. Harrop2 used a value, 609, from a formula given by
Richards. TschernobaefP determined a value of 112 Cals. per
kg. of pure kaolinite. Mellor and Holdcroft,4 by graphical cal-
culation from lag curves, found 42 Cals. per kg. of clay for the
endothermic reaction at dehydration and 21.5 Cals. for the exo-
thermic reaction at about 10000 C. From lag curves given by
Brown and Montgomery,5 approximate calculations gave values
ranging from 757 to 1015 Cals. per kg. of water for clays having
10.7 to 13.2 per cent chemical water. As would be expected,
the heat required increases with the amount of chemical water
in the clay. On the basis of these data the value for the clay in
this test, containing 4.65 per cent chemical water, was taken
as 220 Cals. per kg. of chemical water.

The heat required to raise the temperature of the clay in the
respective chambers was:

1 A. V. Bleininger, Trans. Am. Ceram. Soc., 10, 417 (1908).

2 C. B. Harrop, Ibid., 14, 227 (1912).

3 D. Tschernoboeff, Rev metal., 1905, p. 729.

4 Mellor and Holdcroft, Trans. Eng. Ceram. Soc., 10, 101.

6 Brown and Montgomery, Trans. Am. Ceram. Soc, 14, 711 (1912).

AMERICAN CERAMIC SOCIETY. 575

Chamber 10. 53,320 X 2.5 X 0.23 X 210 = 6,438,000

Chamber n. 54,560 X 2 .5 X 0.23 X 300 = 9,412,000

Chamber 12. 54,560 X 2 .5 X 0.23 X 180 = 5,649,000

Chamber 13. 52,080 X 2.5 X 0.20 X 300 = 9,832,000

Heat for clay = 31,331,000 Cals.

The mechanically held water was expelled in Chamber 13,
squiring 52,080 X 2.5 X 0.014 X 631.2 = 1,073,000 Cals. in
le steam at the flue-gas temperature.

The chemical water was expelled in Chamber 12. The heat
squired for dissociation was 220 Cals. per kg. of water. The
eat in the steam at the flue-gas temperature was 631.2 Cals.
'he heat requirement was then 851.2 Cals. per kg. of water,
ut the ware in Chamber 12 was at an average temperature of
30° C at the beginning of the test and the chemical water in the
ay at the time was assumed to have held the heat equivalent
) that of steam at the same temperature, 715 Cals. per kg. The
ctual amount of heat required for the chemical water was, hence,
36 Cals. per kg.

54,560 X 2.5 X 0.0465 X 136 = 863,000 Cals.
he total heat required for the ware during the test was
3,267,000 Cals.

Heat Input for Kiln. — The heat value of the hot producer-gas,
le heat in the atmospheric air supplied to the kiln, and the heat
om combustion of carbon in the clay constituted the heat input.

The heat in the producer-gas was 9371 X 6086 = 57,032,000
als.

The heat content of the atmospheric air, as supplied to the
roducers, was 44.4 Cals. per 3.79 cu. m. of dry air per kg. of
sal or 11.72 Cals. per cu. m., including the heat in the moisture,
he air supplied to the kiln furnished 411,300 X 11.72 =
,819,000 Cals.

The heat from the combustion of 0.25 per cent of carbon in
le clay in Chamber 12 furnished 54,560 X 2.5 X 0.0025 X

100 = 2,762,000 Cals.

The total heat input for the kiln alone was 64,613,000 Cals., of
hich 33,267,000 Cals. were used for burning the ware (51,49
2r cent), 26,690,000 Cals. were lost in the flue-gas (38 . 2 1 per cent)

576

JOURNAL OF THE

and the balance, 6,656,000 Cals. (10.30 per cent) wen- charged
to conduction, radiation and other undetermined losses.

Considering the kiln and producers as a unit, the total heat
input from coal fired, air supplied to producers and kiln, steam
supplied to the producers and carbon in the clay was 76,318,000
Cals. Of this amount, 43.59 per cent went to the burning of
the ware, 32.35 per cent to the flue-gas, 8.72 per cent to heat
losses from the kiln, 14.72 per cent to heat losses from the pro-
ducers and o . 62 per cent loss was due to unburned carbon in the
ashes. The data for the combined unit, together with the heat

Fig. i.
Heat Balance. Chart.

Heat Balance or G*s Producers.

Toto/ heat input from coo/, oir.
and steam tOO %

Heat /n producer j?as, cafor/f/c
t/o/ue p/uj sensib/e heat 82 97

Loss doe to conduct/on.
radiation, etc,

/OJS

Loss atae fo unhurned carbon
in ashes C.68 Y<.

I otat hear input from producer
J/as, arr ond carbon in c/ay tOO"

fteat fo ra/se temp, of c/ay and
Ctnyt off wafer Jt.49 %

Loss due to sens/6/e
of f/ue j? as

tiear

33 2/ %

Loss due to conducfi on
radsat/on, etc tO.30 °fo

Heat Balance or ff/LN

ffsAT Balance or Kiln ano Producers Combined

Totat A eat /nput from coat, air, 1
Steam and carhon in c/ay /OOfo^^

Heat to raise temp ot c/ay a.^a^^k
o/r/se off water <f3.S9%^k

Loss c/ue to sens'£/e neat H
of f/ue -j>as. 32.JS/" 1

Loss due to conduction roa/ia-^^
tiony etc from Mi In 872 /Z M

Loss due to conduct/on, rad>a - H
tion, etc from producers /-J2/^ ■* BlB

Loss c/ue to uohurned 1
carfion /n ashes ■ 0.62 /« \

AMERICAN CERAMIC SOCIETY. 577

distribution for producers and kiln separately, appear on the
accompanying chart.

Discussion of Results.

The data, as presented, serve as a means of comparing the heat
distribution in a kiln of this type with that in periodic kilns used
for similar products. As would be expected, the losses to be
charged to conduction and radiation are much reduced, since a
considerable amount of the heat going to the kiln structure dur-
ing the firing period was recovered during the cooling — preheating
the air for combustion. Hence, the percentage of the heat input
which may be charged to the burning operation is much increased
over that of the periodic kiln.

Of the heat charged to the heating up of the clay, the greatest
portion was actually lost from the kiln by radiation and conduc-
tion and in the heat in outgoing ware. However, since it was
first applied to the ware to accomplish the burning operation,
it was properly charged thus.

Although the temperature of the outgoing flue-gas was very
low, the large excess of air admitted to the kiln resulted in a large
volume of flue-gas and, consequently, a large loss of heat in the
flue-gas. If some means of regulating the amount of air admitted
were applied, this loss could be materially reduced.

It would be of interest to determine the actual amount of heat
lost from the kiln structure itself and that remaining in the
burned ware and which escaped to the atmosphere by circulation
of the air through the cooling chambers and through the openings
in the crown and the doorways. Considerable difficulty would
have been encountered in obtaining the necessary data for this
purpose.

In conclusion, the writer desires to express his indebtedness to
Prof. C. W. Parmelee for his able assistance in carrying out the
tests, to Mr. C. W. Parks who assisted in obtaining some of the
data, and to the Baker Clay Company and the International
Clay Machinery Company for their cooperation in the work.

Department of Ceramic Engineering,
University of Illinios.

SOME FACTORS INFLUENCING THE TIME OF SET OF
CALCINED GYPSUM.1

Hy !•'. I". HOUSBHOLDSR.

Introduction.

The time of of set calcined gypsum is one of its very important
properties. Several methods for its measurement have been
suggested and the chief ones have been tested and rated according
to their merits.2 However, no data resulting from thorough work
on the relation of the time of set to the other physical properties
is available. Obviously, this relation will depend upon the purity
of the gypsum from which the plaster is made, the temperature
at which it is calcined, and the amount of retarder or accelerator
added. Experience shows that it also depends upon the cleanli-
ness of the boxes or vessels in which the plaster is mixed, and the
purity of the water used.

This preliminary note describes some tests made in order to
determine the effect of varying the consistency of the mixtures,
the time of stirring, the rate of stirring, and the temperature of
the water used in mixing, on the time of set of calcined gypsum.
The tests reported — although they may be of some value in
themselves — were made as a guide in outlining further and more
extensive tests on the physical properties of gypsum.

Preparation of Samples. — The same grade of plaster was used
(without the addition of any retarder or accelerator) throughout
the whole series of tests. A sample of 300 g. was made up in j
each case in a clean casserole. Each sample was stirred by means |
of a small spatula — kept well polished. The plaster was added
to the water as rapidly as possible (4 to 5 seconds) without splash- '
irig or waste and allowed to stand undisturbed for one minute.
It was then poured (or was stirred as the test demanded and then
poured) into the standard mold of a Vicat3 apparatus. The

1 By permission of the Director, Bureau of Standards.

2 Emley, Trans. Am. Cetam. Soc, 19, 573-584.

3 1915 Year Book, A. S. T. M., p. 359.

AMERICAN CERAMIC SOCIETY.

579

iold was thoroughly cleaned before each test and covered with
very thin coating of oil. The time of set was determined, as
lggested by Emley,1 by means of the standard Vicat needle.
Effect of Per cent Water Added. — To determine the effect of
ie amount of water used in mixing the plaster, on the time of
:t, samples containing 30, 321/2, 35, 37V2, 40. 4272, 45» 47V2,
) and 60 per cent of water, respectively, were made up and
:sted. A large number of tests on each mixture were made
id che average results at each percentage plotted. In Fig. 1
shown the variation in the time of set with the amount of water

^ i0

F/G./.

0

1

^

IS J

0 3

S 4

0 4

s s

0 s

s

sed in the mixtures. The samples containing 30 per cent water
ere so stiff and viscous that much care and effort were required
i introducing them into the mold. However, the time of set
F the individual samples did not vary from the average by more
lan one-half minute. The sample containing 32 V2 per cent
ater was too viscous to mold satisfactorily, but gave fairly uni-
>rm values.

The curve (Fig. 1) shows a fairly uniform increase in the time
f set as the percentage of water was increased until about 45
er cent was reached. Then the increase in time of set was less
ipid — due to the fact that when the mixture was poured into
1 See page 578, Loc cit., 19, 583-4.

58o

J( lURNAL OB THE

the mold the plaster settled to the bottom and the excess water
came to the top. A thin him of water came to the top of the
47V2 l)(-'r Cenl sample, and in the ease of the 50 per cent sample,
a layer of water 1 1.5 mm. deep formed, while with the 60 per
cent sample, the water was 7 mm. deep at the top of the mold.
Evidently, before any setting had taken place, the mixture was
reduced to about 47 or 48 per cent water and the further retarda-
tion was probably due, in part at least, to the fact that the plastel
was under water. Also, the layer of water at the top of the mold
increased the penetrating ability of the needle. These samples
did not harden at the top — not even after the pouring o!T of the
water and their removal from the mold. A day or two later the top
could be brushed away with the hand. Considering the samples
containing from 30 to 45 per cent water, an increase in the time
of set from 6.5 minutes to 20 minutes, or approximately 300 per
cent, is noted, while in the 45 to 60 per cent samples it increases
from 20 minutes to 29 minutes or approximately 50 per cent.

Viscosity Measurements. — The relative viscosities of each of
the samples was determined by means of the Southard viscosi-

1S -to £S

EERCEh T- Wfl TER

AMERICAN CERAMIC SOCIETY.

58i

leter. The samples from 35 per cent up were placed in the vis-
jsimeter and gave the results shown by the dotted curve (Fig.
). Assuming that the "patties" were circular, the viscosity
i a mixture would vary inversely as the square of the diameter
: the "patty." The relative viscosities are shown by the solid
irve (Fig. 2). A very marked decrease is noted between the water
)ntents of 35 and 40 per cent but not much change is noted
lereafter. The instrument used was not water tight and hence
le 60 per cent sample was not tested. The shape of the vis-
)sity curve would indicate that the viscosity very rapidly ap-
roaches that of water and probably reaches it before it crosses
le 100 per cent abscissa. Lack of data makes it unsafe to judge
here the two curves meet.

No extensive work has been published on the properties of
fpsum solutions, although investigations along this line may lead
> a more satisfactory definition of normal consistency. The
nmediate results obtained here indicate, that decreasing the
iscosity increases the time of set, but that the relation between
le two is not a simple function.

Effect of Stirring. — In determining the effect of continued
irring, the samples used were mixed with 40 per cent water and
ich was stirred with a small spatula at the rate of four strokes
B second. The calcined gypsum was added to the water and
le mixture allowed to stand one minute, as before. The first
imple was then poured immediately into the mold and the time
' set determined. Each of the others was stirred at the same
ites for o, 0.5, 1,2,3, and 4 minutes, respectively, and then poured
to the mold. In Fig. 3 is shown the results of the tests. Con-
nuous stirring for four minutes decreases the time of set from

o

J {

FIG 3.

>

TIME OF STIRRING-MINUTES

^

Jl »i RNAL OF THE

[8.5 minutes (with no stirring) to 9.5 minutes or to almost one
half. The general shape of the curve would indicate thai the

time of set would not be very greatly decreased by a continual ion
of the stirring.

In Fig. 4 is shown the effect of varying the speed, of stirring of
the same kind of samples on the time of set. The stirring was
all done by hand and the speed varied from o to 6 strokes per
second (each sample being stirred for one minute only). In-

F/G. 4.

^~"~- — ^

1

ftfiTE OF 5TIRRU16 - BEfiTS PER 5ECQHD

creased speed of stirring reduced the time of set. The shape of
this curve is very similar to that of the curve shown in Fig. 3,
although the decrease in the time of set is not so marked. The
increased agitation of the mixture — either by increased stirring
or by more rapid stirring — hastened the time of set. The first
effect is probably to increase the rate at which the dry plaster
takes up the water and goes into solution. The first crystal
clusters formed are probably broken up and distributed through-
out the mixture — thus becoming nuclei for further crystalliza-
tion.

Effect of Temperature. — The effect of heating the water before
mixing with the plaster was also tried out. The samples used
for this purpose contained 40 per cent water and each was stirred
at the same rate and for the same time as before. The data se-
cured is less uniform than that secured in the foregoing experi-
ments, but does not vary in any given direction. The conclusion
reached is, that the temperature of the water has no appreciable

AMERICAN CERAMIC SOCIETY.

583

feet on the time of set. The setting of calcined gypsum is
ways accompanied by an evolution of heat (pure CaS04 V2H2O
forming CaS042H20 giving up about 3700 cal. per gram molecu-
r weight). Heating would therefore tend to retard the process
ere it kept confined. On the other hand, the solubility of cal-
ned gypsum tends to increase with a rise in temperature of the
ater — which would probably induce an acceleration of the pro-
ss. Evidently, from the results of this test, the two factors
entioned would about balance each other.

F/6 5

.0

0

0-

-0

-XI

■3- (

_ -<

<■ 0

_a

--Q

0

0

0

6

<

1

TEnPEFfflTURE OF WATER °C

By the heavy curve (Fig. 5) is shown the time of set as de-
rmined by this test. The dotted curve shows the points at
hich the needle did not penetrate at all, or the time of hardening
the gypsum. The same regularity that is seen in the time of
t is shown here.

Conclusions.

1. Increasing the proportion of water causes a decrease in
scosity and an increase in the time of set of calcined gypsum.

2. Vigorous and continuous agitation, when mixing the calcined
/psum with the water, tends to decrease the time of set.

3. The temperature of the water used in mixing has no in-
aence on the time of set.

The above experiments were the outcome of the suggestions of
tr. C. A. Birdsley, Chief Engineer; P. H. Butler, Supt.; and J. A.
•avis, of the Oakland Plant of the U. S. Gypsum Co. Acknowl-
Igments are also due Mr. F. A. Kirkpatrick for many valuable
iggestions in making the experiments.

THE NATURE OF THE AIR CONTENT OF PUGGED CLAYS.

By II. SPUSKISK.

As a sequel to some work done on the quantitative measure-
ment of the air content of pugged clays, the question arose as to
the composition of such air after imprisonment for a greater or
less period in the clay body. A preliminary analysis of the libera-
ted and collected gases gave an oxygen content of 7 per cent as
against a normal oxygen content in free air of about 21 per cent.

At the time of this determination, it was not possible to secure
a large enough sample to make a complete analysis that could be
relied upon as thoroughly trustworthy. Recently, a very simple
scheme (Fig. 1) was devised to secure any desired amount of the
gas occluded in clay.

A large glass funnel was inverted and equipped with a tube of
30 millimeters internal diameter, suitably drawn out at both

Fig. i.

AMERICAN CERAMIC SOCIETY. 585

ends. The upper end was bent at a right angle and a gas-tight
joint was made at the lower end between the tube and the in-
verted funnel. The funnel was then partly immersed in water
contained in a large iron kettle. The water was boiled for a con-
siderable time to eliminate dissolved gases, allowing steam to
issue at the upper end for about 15 minutes. The screw-cock at
"A" was then closed and the flame lowered till the water was just
at the boiling point. The tube "B," on cooling, filled itself com-
pletely with water — proving the expulsion of the air.

Clay blanks which had been pugged about 24 hours previously
were then slipped under the funnel; as the blanks disintegrated,
the contained air was liberated — rising to the top of the tube "B."
This was continued until a sufficient quantity of the gases had
been collected.

The transfer and analysis of the gases needs no description.
The following analysis represents the collected gases:

Per cent.

Carbon dioxide 3 .85

Oxygen 1346

Carbon monoxide 1 .92

Nitrogen t . . . 80.77

As is well known, if a bituminous coal is kept in a confined
atmosphere, it will extract nearly all of the oxygen therefrom.
As no further work has as yet been done, it would be somewhat
hazardous to draw any conclusions as to what might be called the
secondary effects of the change made obvious by the analysis.
The aim of the present note is to bring the matter to your at-
tention and to stimulate other workers in this direction. There
are indications that the change is not quite so simple as at first
sight might seem apparent, and that further investigation of the
phenomenon might not be barren of results of practical importance.

It is obvious, of course, that these chemical changes are inti-
mately connected with the well-known physical changes that take
place during the ageing and weathering of clays. Some light
might, in this way, be thrown upon the much studied and as yet
imperfectly understood question of plasticity.

Jeffery-Dewitt Co.,
Detroit, Mich.

ACTIVITIES OF THE SOCIETY.

This department has been established by the Publication Com-
mittee, at the request of the President, in order to keep the mem-
bership in general informed of the activities of the officers, the
Board of Trustees, and the several committees of the Society.
This will be the official avenue of appeal to the membership for
cooperation in the work of the Board and committees. Re-
ports of the various Local and Student Sections will be published
here and also accessions to membership, notice of death of members
and other items of Society news. No personal items, unless they
are also distinctly Society news, will be printed.

Next Annual Meeting.

The Board of Trustees has voted that the next annual meeting
shall be held at the Fort Pitt Hotel, Pittsburgh, Pa., on February
3rd, 4th and 5th. Details will be published later.

Resolutions Passed by the Board.

Professional Divisions may be formed in the following ways:

1. "When the initiative in the formation of Divisions is taken by members
who are interested, a petition may be presented by not less than ten mem-
bers in good standing, of whom three or more shall be Active, who are in-
terested especially in some phase of ceramic work sufficiently broad to warrant
the formation of a special Division. This petition shall go to the President
who shall then appoint a representative committee to consider the advisability
of forming such a Division, and to proceed with the organization if the de-
cision is favorable."

2. "When the intiative is taken by the Board of Trustees in order to stimu-
late the growth of the Society, the President shall appoint a representative
man to furnish the initiative. He shall select his own committee."

The Board has passed a resolution to the effect:

"That fifty copies of each issue of the Journal shall be assigned to the
order of the Chairman of the Membership Committee, the number to be sub-
ject to reduction by the Board if the results do not seem to justify the ex-
pense."

AMERICAN CERAMIC SOCIETY. 587

Motions Pending.
Motion by Mr. Montgomery, for immediate decision:

"That the President shall appoint a representative individual for each Di-
;ion to be organized, who shall select his own committee as provided in
ction 2, of the motion recently passed by the Board for this purpose, and
it these committees shall proceed to organize the following Divisions
fore our meeting in February :

1. Glass Division.

2. Enamel Division.

3. Pottery or White Ware Division, including porcelain.

4. Refractories Division.

5. Structural Clay Products Division.

With motion, by the President, to amend by striking out

5. Structural Clay Products Division

and adding:
* 5. Brick and Tile Division, including brick, hollow-tile and drain-tile.

6. Terra Cotta and Faience Division.

7. Abrasive Division."

Motions by the President, for decision December 20th.

1. "That the Committee on Publications be instructed to complete as soon
possible the "Index of the Transactions of the American Ceramic Society"
d to submit to the Board of Trustees an estimate of the cost of publishing
me."

2. "That no Local Section, Student Section or Division shall have the right
assess Section or Division dues against members of the parent Society,

Jess such members have signified their desire in writing to be considered
;mbers of said Section or Division."

3. "That, as of the date of November 1, 191 8, fifty (50) copies of each
lume of the Transactions be reserved for sale in complete sets only, and
at the Secretary be instructed to purchase, at five dollars ($5.00) each,
fficient copies to bring the total stock of each volume to that figure."

4. "That two hundred (200) copies of each number of the Journal be reserved
the time of issue for sale in complete sets only, and that an additional

ro hundred (200) copies be likewise reserved for sale in complete volumes
;ly. Futhermore, that it is the sense of this resolution that a 'complete
I shall consist of the first five volumes of the Journal and such additional
ilumes as shall have been issued up to the date of the sale, and that a 'sale'
all be the actual selling of volumes and not the furnishing of them in con-
leration of payment of membership dues."

5. "That copies of the volumes of the Transactions and of the Journal shall
: furnished to members who have or shall allow their dues to become de-
iquent only after the number reserved for sale has been provided. Further-
ore, that in cases in which the volume that should be furnished in considera-

JOURNAL OF THE

tiou of a certain year's ilm-s is not available, the member shall have tl
option of selecting any available volume or having the dues apply to tl
succeeding year."

Pull discussion by the membership and suggestions in regal

to these resolutions is desired by the Hoard. Any eommunica
tions sent to the President or Secretary will be transmitted t
the other members of the Board. They will not be publish!
here. The President desires suggestions as to proper person
to appoint for the organization of the various Divisions.

ACQUISITION OF NEW MEMBERS DURING NOVEMBER

1918.

Associate.

Frederick K. Minor, Celite Products Co., Pittsburgh, Pa.
John W. Yuill, Hercules Powder Co., Nitro, W. Va.
G. F. Dufour, Y. M. C. A., Hamilton, Ohio.
Melville Marks, Moore & Munger, New York, N. Y.
Dr. A. Bigot, 112 Avenue de Suffren, Paris, France.

Contributing.
Champion Ignition Company, Flint, Michigan.

LOCAL SECTION MEETINGS.

NORTHERN OHIO SECTION.

A joint meeting of the Northern Ohio Section of the America
Ceramic Society and the Cleveland Section of the America
Chemical Society was held in the Hotel Olmstead, Cleveland
Ohio, on December 9, 191 8. The following program was afforde
at the afternoon and evening sessions:

AMERICAN CERAMIC SOCIETY. 589

Ceramic Section — Afternoon Meeting — 3.00 p.m.

"Ceramic Features of the Chemical Exposition," by W. M. Clark, National
imp Works, Cleveland, Ohio.

"Observations on the Apparent Cause of Failure of Lead-Glass Pots," by
F. Gorton, Buckeye Clay Pot Co., Toledo, Ohio.

Discussion — "After-War Adjustments of the Ceramic Industries."

Business Session — By-laws and other important matters relative to the
>cal Sections.

Chemical and Ceramic Sections — Joint Meeting — 8.00 p.m.

"Refractory Materials as a Field for Research," by Dr. E. W. Washburn,
itional Research Council, Washington, D. C.

The meeting was a successful one in every way. It was better

tended and even more interesting than the previous meetings

tiich the Northern Ohio Section has held.

The paper by Mr. W. M. Clark was a very able review of the

eramic features of the Chemical Exposition together with

rtain exceedingly interesting conclusions which Mr. Clark drew

ierefrom.

The paper by Dr. Gorton on "Causes of Failure of Lead-Glass

3ts" was very instructive and promises to be of interest to

aders of the Journal.

Mr. Stowe, in a paper on "Magnesite" (not mentioned in the

rmal program), ably discussed the recent developments in this

lportant branch of the refractories industry.

The paper by Dr. Washburn at the joint meeting in the eve-

ng was a very interesting presentation of some of his impres-

ans and findings as a member of the National Research Council.

his paper will be published in an early number of the Journal.

R. D. Landrum, Secretary, B. A. Rice, Secretary,

Cleveland .Section, Northern Ohio Section,

American Chemical Society. American Ceramic Society.

590 JOURNAL OF TIIIC

NEW JERSEY CLAY WORKERS' ASSOCIATION.

A well attended meeting of the New Jersey Clay Workers
Association was held at Rutgers College, New Brunswick, N. J.,
on December 17th. The following program was afforded at the
morning and afternoon sessions:

Morning Session.

"Some Defects in Terra Cotta," by D. F. Albery, Federal Terra Cotta Co.,
Woodbridge, New Jersey.

"Sagger Mixtures," by Prof. G. H. Brown, Department of Ceramics,
Rutgers College, New Brunswick, New Jersey.

"The Effect of Some Fluxes in Porcelain," by F. H. Riddle, Bureau of
Standards, Pittsburgh, Pa.

Afternoon Session.

"Fuel Combustion," by Henry Kreisinger, Fuel Section of the U. S.
Bureau of Mines, Pittsburgh, Pa.

"Future Kiln Developments," by Prof. Carl B. Harrop, Department of
Ceramic Engineering, Ohio State University, Columbus, Ohio.

"Solving One of the Clay Manufacturers Greatest Problems," by Mr.
Paul D. March, Philadelphia Textile Machinery Co., Philadelphia, Pa.

The New Jersey Clay Workers Association will apply for a
charter as a Local Section of the American Ceramic Society
at the annual meeting of the Society in February.

JOURNAL

OF THE

AMERICAN CERAMIC SOCIETY

A monthly journal devoted to the arts and sciences related to
>e silicate industries.

d1. 1 September, 1918 No. 9

EDITORIALS.

THE PATENT SITUATION.

A mode of relaxation available to any ceramist is the study of
irrent patent specifications referring to inventions of a ceramic
iture and which is certain to afford a constant source of amuse-
ent. Still, a feeling of apprehension is apt to creep over the
isophisticated follower of the fictile arts when he discovers that
i is daily and unsuspectingly violating valuable privileges by
aking use of the common constituents of his products. For, have
)t patents been issued upon the employment of kaolin, feldspar
id quartz in ceramic bodies? Indeed, this is not a jest but the
>lemn truth as may be readily proven by reference to the official
sts. And what of the weird compositions of glasses, enamels
id glazes to say nothing of the wonderful compositions of refrac-
>ries, cement brick and other products which have been duly
rotected. Equally elevating is the perusal of the patent descrip-
ons referring to kilns, driers, furnaces and other appliances
nployed in our art, some of which represent ideas hoary with
itiquity.

Seriously speaking, the situation as regards ceramic patents is
ppalling. It is evident that our whole system of affording patent
rotection to inventions relating to our industries is without
lyme or reason. Like heaven's mercy, patents fall both upon the
ist and the unjust and the result is injustice and chaos.

So much has already been written on the general deficiencies of
ur patent system that it seems futile to join in the chorus, but
le situation is so grotesque when applied to the silicate industries
lat it is our duty to raise our voice in protest.

592 JOURNAL OF THE

Such a complaint is 1 1 it- more justified when one considers that
the remedy is neither difficult nor complex. The creation of a
small committee of experts for each of the natural divisions of
industry and technology, among them one for the field of ceramics,
would bring about the desired change, namely, full protection of
original ideas and rejection of the undeserving claims. By mak-
ing the experts responsible to the Government in their oath of
office, there is every reason to believe that under the guidance of
the regular officials the work would be well and efficiently done.

It would be folly to place the blame for the present conditions
upon the hard-working examiners of the patent office. It was not
intended that they should be omniscient and they are doing the
best they can. However, when so many things are being changed
for the better these days it is our devout hope that something be
done in the matter of our patent system.

PROFESSIONAL DIVISIONS.

It would appear that in the "Professional and Industrial Divi-
sion" idea, there lie great possibilities for a broader and more
useful activity and permanent growth of the American Ceramic
Society. Carefully and successfully worked out, it will remove
the greatest barrier to growth in membership with which the
Society has had to contend in the past, namely: indifference
of members, especially the practical man, to regularly attenc
meetings at considerable expenditure of time and money whei
90 per cent of the program is either "high brow stuff" passinj
wray over his head, or else devoted to an industry having nothinj
akin to his own.

In order to secure the fullest exchange of views on mutu£
problems, the confreres must feel at ease, which is certainly nc
the case when the average man gets on his feet to discuss a pape
in a regular meeting, with the stenographer interrupting in aboi 4
the middle of the second sentence to ask his name and the mal
himself realizing that if he should happen to use the word "aboull
or "approximately" someone would be sure to pop up tl ]
minute he was through and "object to the use of the word 'aboul
on the floor of this Society, etc." Everyone who has ever attendi I

AMERICAN CERAMIC SOCIETY. 593

le Society meetings knows that time and again points are raised
hich certain men in the room could and would gladly clear up
ere it not for a reluctance they feel in rising to their feet under
le conditions which usually prevail at our general meetings.

Industrial Divisions will do away with this to a large extent,
[embers in the same or allied industries will gain the confidence
i and learn to know each other better and great good will result,
f course, it is not to be presumed that the general meetings will be
itirely superseded by the Division meetings, but rather that

portion of the time will be set aside for the latter purpose.
: the Industrial Divisions could continue to hold periodic meetings
iroughout the year, so much the better, but even if the Division
leetings are confined to two each year, at the annual and summer
leetings, they will justify their organization.

THE GLASS INDUSTRY.

The desirability of securing the more active cooperation of the
lass industry with the American Ceramic Society has been re-
eatedly emphasized. The rapid progress which has been made
1 this country during the past four years in the technical ad-
ancement of certain branches of this industry should serve to
svaken the interest of the other branches in a representative
jchnical organization such as the American Ceramic Society.

The Society has been fortunate in securing a gratifying num-
er of excellent papers on glass for publication in the numbers of
le Journal issued thus far and in receiving the assurance of an
dequate supply of papers of interest to glass technologists for
le future numbers.

Particular attention is called to the fact that the present num-
er includes three excellent papers of both practical and technical
alue to those engaged in the glass industry. It is to be hoped
hat our members will bring this number to the attention of their
riends and acquaintances who have not as yet become actively
iterested in the Societv.

ORIGINAL PAPERS AND DISCUSSIONS.

THE IDENTIFICATION OF "STONES" IN GLASS.

By N. I.. Bovvun.

Introduction.

In glasses intended for a great variety of uses, and particularly
those for use in optical instruments, homogeneity is a principal
requirement. It is one of the tasks of the glass-maker, therefore,
to eliminate foreign particles and inhomogeneity of any kind in
such glasses. Foreign particles may be of a gaseous nature, when
they are termed seeds, bubbles, blisters, etc. Foreign material
may be of a vitreous nature, when it is described as tears, cords,
striae, etc., and, finally, foreign particles of a crystalline nature
are commonly called stones. Complete elimination of all of these
is never attained in glass practice, with the result that selection
of the product is usually necessary and the amount of acceptable
material becomes a measure of the efficiency of the process of
manufacture. In this paper attention is directed towards the
occurrence of stones in glass, and particularly towards the infor
mation regarding the nature of stones that is to be gained througr
the use of the petrographic microscope. Since an exact knowledge
of the nature of a defect in any material may be a fundamenta
factor in achieving its elimination, it is hoped that this study o:
stones in glass may be of some service to the glass-maker. Ill
should be remarked that most of the stones examined by the write I
were those encountered while he was engaged in the production ol
optical glass.

Varieties of Stones.

It has been the experience of glass-makers that a variety ol
factors may give rise to stones. A classification of stones accord I
ing to their origin may be given as follows:

AMERICAN CERAMIC SOCIETY. 595

1. Pot stones

2. Batch stones

3. Crown drops

4. Devitrification stones

It is often a matter of great importance to a glass-maker, when
; glass is running stony, to be able to refer the stones definitely-
one of these four classes. If they are pot stones, *". e., actual
,gments of the pot, he must look for the trouble in his clays,
his treatment of the pots, in excessive temperature of his fur-
ces or in a particularly, and perhaps unnecessarily, corrosive
.ss batch. If they are batch stones, or residual batch material
dissolved, he must look to the grinding and mixing of his batch
iterials, to a proper balance in the proportions of these materials,
to the possibility of too low a temperature at some stage of the
lting process. If they are crown drops the grade of bricks
:d in the crown or the length of flame must be considered. If
:y are devitrification stones, formed by devitrification or crys-
lization of the glass after it has been completely melted, he must
k to the rate of cooling or to an ill-chosen glass composition.

Identification of Stones.

Chemical analysis of stones is sometimes resorted to in order
determine what their origin may be, but it presents consider-
e difficulty. Stones are usually quite small so that the chemist
st, as a rule, work with a small amount of material; indeed it
;ometimes impossible to obtain an adequate amount. It is no
y matter to separate stones from the enclosing glass, so that he
st work with contaminated material. As a result of these
ious factors a chemical examination may consume much time
1 even then may be unsatisfactory. Then again there are
tain types of stones whose chemical composition alone may
ve one in some doubt as to the class in which they should be
ced. With the aid of the petrographic microscope, on the
er hand, it is usually possible to determine rapidly and deci-
i\y the origin of a stone, since not only its mineral composition
leterminable, but also its internal structure and its relations
the surrounding glass.
Ahe best method of examining stones under the microscope is

596 JOURNAL OF THE

to crush them t<> a powder and examine the powder in a drop o
liquid on a slide alter covering with a cover slip. With the ai<
of a set of liquids of various known refractive indices, the high anc
low indices of the crystals can he measured by the method of com
parison and at the same time extinction angles, optic axial angle;
and other distinctive optical properties observed. The met hoc
is highly advantageous for fine materials intimately intergrown
because immersion in a liquid corresponding in index to one o
the materials presents the other in marked relief and renders pos
sible a study of the details of the manner of intergrowth.

Pot Stones.

Pot stones are occasionally large enough that when broken opei
they reveal their nature to the unaided eye. Usually they an
quite small, however, and even when formed from pot materia
that burns very dark they are snow-white and in this respec
indistinguishable from stones of different origin. Under thl
microscope, pot stones are found to be made up, in all cases, o:
crystals of sillimanite with a little interstitial glass. The silli
manite or aluminium silicate (Al2Si05) is recognized by its usua
crystallographic and optical properties identical with those of th»
natural mineral. Sillimanite is never found in stones of othe
origin.

It may be remarked in passing that this recrystallizing actio
of the glass on the pot material is operative not merely on fraj
ments of the pot that have become detached, but also on the wa
of the pot itself. The result in this case is the formation of a thi
but conspicuous white layer on the inner surface of the pot that
made up of minute crystals of sillimanite with a little interstiti
glass.1 In attacking pot material the glass evidently acts sele
tively, dissolving first from the clays the portion of the silica
excess of that necessary to form sillimanite. All glasses act alii
in this respect so that sillimanite is a universal constituent
pot stones and of the white layer formed on pots whatever mi
be the composition of the glass. As a result of this selecti
action of glass on clay material, the more silicious pots are mu
1 G. V. Wilson has examined this white layer and found it to eonfc
crystals of sillimanite. See S. N. Jenkinson, J. Soc. Glass Tech., 2, No. 5,
(1918).

AMERICAN CERAMIC SOCIETY. 597

ore vigorously attacked than the more aluminous varieties,
lis does not necessarily mean that the silicious pots give more
Dny glass, however, for the stones formed may be more readily
similated by the glass. It is rather a lack of uniformity in pot
aterial that is to be avoided, for this may give local pitting and
e casting of highly resistant stones.

Having ascertained that stones have their origin in his pots,
e glass-maker may then take such steps to remove them as
turally suggest themselves, such as more careful preheating of
; pots, altering the glass batch to render it less corrosive, melt-
s' at lower temperature, and, of course, altering the pot batch.

Batch Stones.

The only normal constituent of a glass batch that is ever left
assimilated by the glass is the sand. It then occurs as whitish
lins embedded in the glass and having the same general appear-
ce as stones of any other type. Under the microscope batch
>nes are found to consist entirely of silica in one of its forms,
equently there persists in the grain a core of the original quartz

the sand and this is surrounded by a layer of cristobalite or a
xture of cristobalite and tridymite. Occasionally a small grain
ly be altered entirely to cristobalite and tridymite, when the
re is a dense mass of that material without interstitial glass,
fering in this respect from these forms of silica when they occur

devitrification stones to be described later. These various
ms of silica are readily recognized by their characteristic op-
al properties. Excessively fine-grained cristobalite and tri-
mite are not readily distinguished the one from the other, but
s is often a matter of no great moment since the material is
idily recognized as one of the low-refracting forms of silica.
A. batch stone is usually surrounded by an aura of glass of lower
ractive index than the glass as a whole and frequently there
a "tail" of similar material (see Fig. 1). These are the result
the gradual solution of the silica of the stone and are wholly
:king in the case of silica stones that originate through devitrifi-
tion.

For the elimination of batch stones, the glass-maker should
dt at a higher temperature or introduce more "fluxes" into the

598

JOURNAL OF THE

batch, that is, he should proceed in a manner directly opposed h

most respects to that required for the elimination of pot stones

wherefore the importance of distinguishing these two types d

-tones. There are, of course, other possible remedies for batd
stones, such as the use of finer or more uniform sand, the carefu
mixing of batch ingredients, etc.

■ •■ ,k,; • 7. ~ >

-

Fig. i — Batch stones with tails. An extreme example.

Occasionally, through the careless handling of batch materials'
they may acquire extraneous matter that is unassimilable by th
glass and therefore appears as stones. Identification of sue
stones may sometimes be accomplished under the microscope an
a clue as to source thereby obtained.

Crown Drops.

When the crown of a furnace becomes rather badly corrodf
by the vapors that are given off by glasses, some liquid may dr
from the bricks of the crown and carry particles of the brick wi'
it. Where silica bricks are used, as they commonly are in Amet
can practice, grains of silica are thus introduced into the met
and may persist in the finished product. At first thought it wou
seem that particles of silica added in this manner would differ
no particular from grains left over from the batch, but as a mati

AMERICAN CERAMIC SOCIETY. 599

fact they usually can be distinguished. When quartz is acted

by fluxing materials it changes first to cristobalite, even at
tnperatures where tridymite is the stable phase, and then the
stobalite is gradually made over into tridymite. For this
ison the silica of batch stones, which is fluxed in the glass for a
mparatively short period, is predominantly cristobalite and
ually very fine grained, for it is continually dissolving and has

opportunity to grow to large dimensions. The silica from the
)wn, however, has been fluxed for a long period in the liquid
at forms on the surface of the bricks before it arrives at that
ite where drops are likely to form. In the writer's experience

both regenerative and blast furnaces, complete conversion to
dymite is accomplished in this surface material and rather large
^stals have grown that are easily distinguished from the fine-
ained cristobalite of batch stones. From furnaces whose crowns
2 maintained at temperatures above 14700 C for lengthy pe-
•ds, which is rarely done in glass practice, cristobalite should be
-med rather than tridymite, and Merwin describes cristobalite
>m the crown of a recuperative furnace.1 In such a case the
le that crown drops are made up of tridymite would not hold,
t even here it seems likely that the cristobalite would have
awn to much larger dimensions than in batch stones and might

distinguished by this fact. Of course one can always deter-
ne, for any particular furnace, the exact nature of the material
at would enter into the formation of crown drops by chipping
' some of the glaze from the bricks of the crown when it is cold
ough to render this possible.

Devitrification Stones.

Devitrification or crystallization of glass may make its appear-
ce in various forms. When crystals are formed that are large
ough to be seen as individuals with the unaided eye and are
sseminated through a glass which is otherwise clear and trans-
ient, they are usually described simply as crystals and the
iss is said to have undergone "crystallization." When the crys-
Is are small and closely spaced, so that the effect is to render

1 Ferguson and Merwin, "The Melting Points of Cristobalite and Tri-
mite," Am. J. Sci., [4] 16, 424 (1918).

600 JOURNAL OF THE

tlu- glass an opaque white miss, the glass is said to have under
gone "devitrification." There is, however, no essential difference
and such difference as appears is simply the result of the limita
tions of the human eye. Under the microscope- a piece- of "de
vitrified" glass is seen to consist of a clear transparent glass base
through which crystals are dotted, the whole appearance being
the same as that presented to the unaided eye by a piece of glass
that has undergone "crystallization," as that term is used above
While the formation of minute crystals in glass, in such mannei
that the whole mass is rendered opaque, can hardly be referred tc
as the formation of stones, it is, however, considered to somt
extent here since the phenomena are fundamentally the same
A common misconception concerning devitrification should, per-
haps, be mentioned at this point. It is sometimes considered
when a piece of glass is rendered a dense opaque mass, that it has
crystallized as a whole. This is far from the truth. Almosl
invariably there are formed crystals of only one of the compounds
occurring in the glass and the crystals themselves seldom amount
to more than 10 or 15 per cent of the total mass of glass.

A common form of crystallization or devitrification is that ir
which spheroidal masses of crystalline material are formed in the
glass. These are analogous to the spherulites that occur in natura
glasses, i. e., volcanic lavas. In artificial glasses they are made
up of innumerable minute crystals of one of the constituents o;
the glass radiating from a center and having glass in the inter
stices between the crystals or radii. These spherulites consti
tute the typical devitrification stones. Occasionally crystalliza
tion of this kind begins with the formation of a rather large indi
vidual crystal and is continued, as the glass becomes more viscous
by the growth of a spherulite about this crystal.

Silica, the fundamental constituent of nearly all glasses, is pel j
haps the commonest ingredient that separates from them. 1
may separate as minute crystals, evenly disseminated, when
gives an opalescent, milky or a dense opaque glass according t
the size of the crystals. Under exceptional conditions large ind
vidual crystals of tridymite have been noted.1 Finally it ma
1 H. Le Chatelier, Bull. Soc. Min. Fr., 39, 150 (1916).

AMERICAN CERAMIC SOCIETY. 601

parate in the form of spherulites which have special interest in
lis connection since they constitute typical stones.
The silica contained in spherulites of this type is, in the writer's
perience, principally in the tridymite form though some eristo-
ilite is usually present and occasionally may predominate,
tiough a spherulite may appear to be entirely crystalline it is
ally composed principally of glass with a relatively small pro-
>rtion of crystals, the glass being interstitial between the radi-
ing crystal fibres. While frequently of the same mineral con-
itution, these spherulites are unlike batch stones in that there
no central core of a dense crystalline nature free from glass,
id there is, of course, never a core of quartz. The interstitial
ass of tridymite spherulites is appreciably higher in refractive
dex than the main mass of glass surrounding the spherulite
:cause there is an abstraction of silica from the interstitial glass,
tie glass surrounding a spherulite is, however, not appreciably
fected by this action. A spherulite grows entirely, then, at the
pense of the glass contained within it. It is for this reason
at chemical analysis of a spherulite shows it to have sensibly the
me composition as the main mass of glass, a fact which has
metimes led to the erroneous conclusion mentioned above, that
ie glass as a whole mass has crystallized.

The compound CaSi03 separates rather frequently from glass,
bis compound is known in two crystalline modifications, wollas-
nite, stable at temperatures below 11700 C, and pseudowollas-
nite, stable at higher temperatures. According to lye Chatelier,
;eudowollastonite forms in some glasses, 1 but the writer has seen
dy wollastonite. In one example seen by the writer, fairly large
dividual crystals up to 3 mm. in length were found in plate
ass. Usually, however, the wollastonite forms typical spherulites.
>me spherulites of this type found in a window glass tank (see Fig.
and examined by the writer, show radiating prismatic fibres of
ollastonite with some interstitial glass that is thickly dotted
ith minute crystals of cristobalite. Wollastonite crystals as
und in glass have properties identical with those of the natural
ineral and given on a later page.

Besides wollastonite, which may separate from ordinary lime-
1 Bull. Soc. Min. Fr., 39, 150 (1916).

6oa

Jl >i RNAL OF THE

soda glasses, and silica, which may separate not only from these

but also from many special glasses, there arc' certain compounds

that separate only from special glasses.

Fig. 2. — Spherulites of Wollastonite (CaSi03) in Glass.

From some barium crown glasses the compound BaSi2Os may
separate.1 It may form fairly large individual crystals (see Fig.
3) if crystallization begins at a high temperature, and about
these crystals spherulitic growth may take place as the tempera-
ture falls. If crystallization takes place entirely at the lower
temperature a simple spherulite may form. Such spherulites or
stones are readily distinguishable from other stones by the char-
acteristic properties of BaSi205.

From glasses very rich in lead, spherulites sometimes separate,
the fibres of which appear to agree in general properties with the
natural mineral alamosite, PbSiC>3.

All spherulites have the same radiating fibrous structure that is
described in connection with silica spherulites.

All of these types of devitrification stones are readily recognized
1 N. L. Bowen, "Crystals of Barium Disilicate in Optical Glass," J-
Wash. Acad. Sci., 8, 265 (1918).

I

AMERICAN CERAMIC SOCIETY. 603

as such by the characters that have been mentioned here in
connection with them. The ability to recognize them may, in
itself, be an important aid to the glass-maker, but the ability to
determine the exact composition of the crystals that separate
may be a still more important factor. The separation of a
certain compound frequently indicates an excess of that compound
in the glass and may point to mistakes made in the batch room or
to variability in the composition of one of the raw materials.
The nature of the compound separating suggests which one of
the raw materials should be suspected. If these factors have

Fig. 3. — -Crystals of BaSi205 in Glass. (Natural size).

been checked and found satisfactory, then the difficulty must be
sought in an ill-chosen batch composition. Nearly always some
readjustment in the chemical composition of a glass can be made
without making a change of properties too great to spoil the
glass for its particular use and the nature of the crystals that
separate furnishes a clue to the direction in which this change of
composition should be made. When a change of composition is
not permissible, the glass-maker must endeavor to vary his heat
treatment in such a way as to reduce the formation of the devitri-
fication stones to a minimum.

Optical Properties of Crystals Occurring in Glass.

The following is not intended to be a complete statement of the
optical properties of the crystals described. It gives merely those

604 JOURNAL OF THE

properties thai arc of greatest use in distinguishing the crystals
as they are found in the glass.

Sillimanite ' \ISi<>,) is orthorhombic and occurs in glass 1U
lath-shaped crystals with parallel extinction and positive elonga-
tion. The refractive indices are 7 = 1.6X1 and a = 1.660 and
can be measured even on very small crystals embedded in glass,
for by careful search of the powder immersed in liquid one
can always find crystals that have not been broken across and
therefore present an edge against the liquid. The refraction is
higher than that of all glasses except the very heavy lead glasses,
which fact enables one to distinguish even excessively minute
crystals from those of cristobalite or tridymite. The double
refraction, 7 — a = 0.021, is moderately strong.

Silica may occur in glass in any of the three forms quartz, tri-
dymite or cristobalite.

Quartz occurs only rarely and then as sand that has remained
unaltered. It is uniaxial and positive; e = 1.553 and w = 1-544-

Tridymite occurs as thin hexagonoid plates that appear as
needles when turned on edge and have parallel extinction and
negative elongation. They frequently show the characteristic
parallel growth of overlapping plates. The refractive indices
are 1.469 -1.473, the refraction and double refraction being
therefore both very low.

Cristobalite frequently occurs as skeleton crystals consisting of
minute octahedra arranged in parallel growth on a cubic pattern
and having a general similarity to magnetite, native gold or other
cubic minerals when they form skeleton crystals. When free to
grow larger it may form elongated cubes that are, however, never
sufficiently elongated to be confused with "needles" of tridymite.
The cubes are sometimes terminated by octahedra that have
somewhat greater dimensions than the cube and give club-shaped
or spear-shaped crystals. The refractive indices are 1.484 -
1.487, somewhat higher than those of tridymite. Probably as a
result of the compensating effect of intricate twinning, cristobalite
usually appears to be nearly isotropic.

Barium disilicaie (BaSi205), is orthorhombic and occurs as thin
six-sided plates (see Fig. 3) with parallel extinction and negative

I

AMERICAN CERAMIC SOCIETY. 605

elongation. The refractive indices are 7 = i.6i7anda = 1.598.1
In connection with the sign of the elongation it should be noted
that a lath-shaped section representing a plate broken across may
be either positive or negative.

Wollastonite (CaSi03), though monoclinic, appears in glass as
crystalline needles or laths with parallel extinction, because elon-
gated parallel to b, the axis of symmetry. The elongation is either
positive or negative, depending on the plane in which the lath is
turned. The refractive indices are 7 = 1.633 and <x = 1.620.

Summary.

The petrographic microscope is a convenient and efficient in-
strument for the determination of the nature and origin of "stones"
or crystalline particles occurring in glass. Stones are divided
into four classes: (1) pot stones, (2) batch stones, (3) crown drops,
(4) devitrification stones. These classes have distinctive features
of structure and texture that are revealed by the microscope.
Moreover, the crystalline phases contained in stones can be iden-
tified by a determination of their optical properties. The results
of a study of stones by these methods are given in this paper.

Geophysical Laboratory,
Carnegie Institution of Washington,
Washington, D. C, December, 1918.

1 For a more complete description of barium disilicate see N. L. Bowen,
J. Wash. Acad. Sci., 8, 265 (191 8).

SOME TYPES OF PORCELAIN.'

By F. H Riddi.k and W. \V. McDanbl, Pittsburgh, Pa.

Introduction.

The present work deals with porcelain bodies burned at tem-
peratures equivalent to the softening point of cone 10 or above.
This includes the firing range of electrical porcelain and of most
hotel china as well as of wares fired to higher temperatures — such
as chemical porcelain and some of the imported chinas. Some
table porcelain made in the United States is burned to cone 13
or 14 but the production is very small. Some of the French,
Austrian and other imported wares are burned at from cone 15
to 18.

The European practice is to bisque the ware at a low tem-
perature and to carry the glost temperature to the maturing
point of the body. This, of course, necessitates the use of a glaze
maturing at the same temperature as the body. The glaze and
body are thus brought into very intimate contact, since the glaze,
as it softens, is absorbed by the porous body before the latter
closes up and becomes impervious. This penetration of the
glaze is merely along the surface but is sufficient to produce a
strong bond between the body and glaze.

The American practice is to mature the body in the bisque
burn and to use a glaze which matures at a lower temperature.
For example, we may bisque to cone 11 and glaze to cone 4 or 5.
This method has many manufacturing advantages. It is not,
however, the purpose of this paper to discuss the relative merits
of the two processes.

The Field Covered.

The chief purpose of this study was to determine the burning
range of porcelain bodies having compositions covering the ordi-
nary porcelain field fired at cone 10 and above. In Table 1
are given the compositions of the bodies used. It will be noted
that the clay content varies from 45 to 85 per cent and that the
1 By permission of the Director, Bureau of Standards.

AMERICAN CERAMIC SOCIETY.

607

Table i.1

Compositions Usee

.

Body
number.

Clay
content
Per cent

Feldspar
content.
Per cent.

Whiting
content.
Per cent.

Flint
content.
Per cent

16 to 23

45

16-30

39-25

24 to 31

50

16-30

34-20

32 to 39

55

14.5-28.5

I

5

29-15

40 to 47

60

14.5-28.5

I

5

24-IO

5 1 to 55

70

IO-28.5

I

5

18.5-O

56 to 59

75

IO-23.5

I

5

I3-5-0

60 to 62

80

IO-18. 5

I

5

8.5-9

63

85

135

I

5

O

ux content varies from 10 to 30 per cent. In several of the
odies of high clay content the flux content has necessarily been

CL/ir

Fig. i.

1 The clay content was made up of equal parts of Georgia, Florida,
tforth Carolina and Delaware kaolins. Where the clay content was over 55
•ercent the balance was calcined to cone 14 and ground dry to pass a 120 mesh
ieve.

608 JOURNAL OF THE

reduced. No ball clay was used in any of the bodies of the first
series, the clay content having been derived from equal parts of
four American clays, namely, North Carolina, Florida, Georgia

and Delaware kaolins. In all cases in which the clay content was
nunc than 55 per cent, the balance was calcined. It would not,
of course, be a common practice in the manufacture of ware,
excepting in a few cases, to entirely eliminate ball clay. The
effect of ball clay is shown later. It is a fact that most of the
vitreous pottery bodies made in this country at the present time
contain but small amounts of ball clay. In the triaxial diagram
(Fig. i) is shown the field covered.

Methods of Testing.

The method of procedure in testing the different bodies was
that ordinarily used in work of this kind. The bodies were
weighed out from standard grade materials — previously dried so
that no allowance was necessary for moisture. Two five-kilo-
gram charges of each body were weighed out and ground wet for
three hours in a ten-inch ball mill. Throughout the work all
conditions of testing were maintained as nearly constant as pos-
sible. After grinding, the two bodies were mixed, passed through
a 140-mesh silk lawn, and filter-pressed. The filter cakes were !
then thoroughly hand-wedged and made into ware. Unless
otherwise noted, the material was made up into ware three days '
after having been filter-pressed. The only bodies that were pre- :
pared under different conditions were those made with varying
times of grinding and with varying times of ageing. The bodies
were molded principally into draw trial pieces and standard cups
for the determination of the dielectric resistance. In certain
eases moduli of rupture bars, crushing cubes, and pieces of chem-
ical- and table-ware were made, this depending upon the character
of the body.

The tests on each body included the determination of the water
of plasticity, the drying shrinkage (by volume), the apparent spe-
cific gravity in the dried state, and the porosity and volume
changes at different firing temperatures. In some cases addi-
tional tests such as the transverse, crushing, etc., were made.

AMERICAN CERAMIC SOCIETY. 609

Water of Plasticity.
The previously prepared clay, in the proper working condition,
iras forced from a hydraulic-plunger press through a die iVs"
quare (with slightly rounded corners). The bars were cut into
est pieces 1V2 inches long. For the determination of the water
f plasticity, three of the pieces were weighed immediately after
ieing made and again after having been thoroughly dried at 1 10 ° C.

Drying Shrinkage.

In making this determination the wet briquettes were immersed
1 kerosene immediately after molding and weighing. The vol-
:mes were then determined in a voluminometer of the Seger type,
provided with a burette reading to 0.05 cc. After removing the
pecimens, they were dried, first at atmospheric temperatures and
nally at no° C. The dry specimens were weighed at once,
gain immersed in kerosene (usually for twelve hours) and the
olumes were then determined as before. The shrinkage was cora-
luted in terms of the dry volume. From the weights of the dry
pecimens and their volumes, the apparent specific gravities were
omputed.

From the wet and dry weights and the wet and dry volumes of
he briquettes and the true specific gravity of the clay, it was
lossible to calculate the volume of the shrinkage and the pore
yater in terms of the true clay volume. These computations
rere based on the following relations :

, _ . _ ( volume of shrinkage water, in terms

^— - = < of the true clay substance, ex-

I pressed in per cent.
, , , ( volume of shrinkage water, in terms

— = < of the true clay substance, expressed

( in per cent.
In these formulas:
S = Specific gravity of body (approaching the value of 2.6 quite

closely)
Vi = weight of briquette in the wet state
V2 = weight of briquette in the dry state
Vj = volume of briquette in the wet state
V2 = volume of briquette in the dry state

610 l< »i 'RN'AI. OF THE

The value a — ■ l> represents, similarly, the volume of por<

</ b
water and the ratio - , the relation of pore to shrinkage wain.

By dividing b by a, and multiplying by ioo, the percentage of
shrinkage in terms of the volume of total water may be computed.

Modulus of Rupture.

Test pieces in the form of bars 7" X 1" X l" were carefully
molded from a thoroughly wedged piece of clay by wire cutting
to as near the mold size as possible and then carefully placing
in the mold in one piece, pressing in, and trimming off with
a wire at the top. The properly numbered pieces were dried at
room temperature, heated at 1100 C, cooled in a desiccator, and
immediately broken transversely. In all cases ten pieces were
broken and the average result taken.

' Burning Behavior of the Clays.

The draw trial briquettes, after having had their dry volumes
determined, were placed in a test kiln, fired with natural gas and com-
pressed air, and burned at a rate corresponding to a temperature
increase of 20 ° C per hour above 800 ° C. Beginning at 12000 C,
the briquettes were drawn from the kiln at intervals of 30 ° C.
This was continued up to 14400 C and, in case the bodies were
extremely refractory, the temperatures were carried higher until
the pieces became overburned. All pieces, as they were drawn
from the muffle, were placed as rapidly as possible in an annealing
furnace and cooled down gradually in order to prevent shattering.
When cool, the briquettes were weighed, immersed in water,
boiled in a partial vacuum for five hours, again weighed suspended
in water, and finally weighed in air when fully saturated with
water.

From these data, the porosity was calculated by the use of the
Purdy formula.1 The volumes of the briquettes were determined
by means of the voluminometer — kerosene being used as the dis-
placing liquid. The volume shrinkage or expansion was invar-'
iably expressed in terms of the original volume in the dried state.
Both the porosity-temperature and the volume-temperature rela-
1 R. C. Purdy, 111. Geol. Surv., Bull. 9.

AMERICAN CERAMIC SOCIETY. 6ir

lions are extremely significant in the study of the burning range
md behavior of a body. When plotted, the information presents
in easy method of study and comparison. Zero porosity, or bet-
ter, zero absorption, may be considered the first point of maturity.
\ more sensitive criterion of maturity is the point of greatest
iensity or smallest volume and it should correspond to the most
suitable temperature for burning the ware. As soon as over-
turning— due to the formation of a vesicular structure — starts,
:he volume increases and a drop in the curve is noted. It is
evident, then, that the greater the difference in temperature
Detween the point at which zero porosity is first reached and that
it which overburning is evident, the wider is the firing range of
i body. When the overburning advances far enough, there will
De an increase in porosity — due to water entering some of the
>utside pores that have been ruptured. The volume change will,
:herefore, indicate overfiring earlier than the porosity relation.

Considerable care and accuracy is necessary in making draw
xial burns in order to make comparison possible between several
)urns. Thus far, seven burns have been made in the work under
liscussion. The kiln was large enough to hold all of the briquettes
or seventeen bodies. As a check against the thermo-couple and
>yrometric cones, it has been our practice to burn a set of draw
rials of the same body in several burns and to check the results
)y comparison.

rhe Effect of Increasing the Feldspar Content and Maintaining
the Clay Content Constant.

This is very well illustrated in Fig. 2. The bodies used were
lumbers 17, 19, 21 and 23 of Table 1, their compositions together
vith their burning ranges being given in Table 2.

Table 2.

Composition

Temp, of

zero
absorption.

Temp, of

maximum

density.

No.

Per cent

clay
content.

Per cent
feldspar
content.

Per cent

flint
content.

Temp, of

start of

overburning.

17

45

18

37

13800 C

14100 C

14400 c +

19

45

22

33

13500 C

13800 C

14100 c +

21

45

26

29

13200 c

1350° C

14100 c—

23

45

30

25

13200 c

1350° C

13800 c

6l2

JOURNAL OF THE

A study of the results shows thai the range is not much
wider for the low feldspar than for the high feldspar bodies —
though there is some difference. This is better shown by the
curves than by the numerical values. The- curves also indicate

/S00 /230 /?£# /2SH7 /320 /JS0 /30O /4/0

TfrtP£/?/?TUft£ /rt-eC.
Fig. 2.

/tie /4ra

that the bodies of high feldspar content appear to approach zero
absorption more abruptly. Although the firing ranges of the
several bodies are nearly the same, the temperatures of maturity
are much lower in the case of the high feldspar bodies — as would
naturally be expected.

This same type of curve applies to bodies having a much higher
content of clay — up to 75 per cent. In Fig. 3 it will be noted
that bodies Nos. 52 and 55 contain 70 per cent clay and that
body No. 52 contains 16V2 per cent flux and body No. 55 contains
30 per cent flux.

The chief difference in the two sets of bodies, as shown by Figs.
2 and 3, is that the variation in the content of flux has a more
marked effect on the low clay (Fig. 2) than on the high clay
bodies (Fig. 3). This would indicate, therefore, that the higher

AMERICAN CERAMIC SOCIETY.

613.

;he clay content, the less the effect of variations in the amount
)f flux. The difference, however, is not great. In fact, it is
ather surprising how little difference there is in the burning

/200 /Z30 /260 /£S>0 /320 /3S<? /300 /4/0 /440 /47V

Tfffp/rff/in/ffjr ///- v

Fig. 3.

anges of bodies having varying clay contents, so long as the
tmount of flux used is constant, the flint content decreasing, of
:ourse, as the clay content increases.

In Fig. 4 is shown the burning curves for the three bodies, the
ompositions of which are given in Table 3.

Table 3-

Body
No.

Per cent
raw
clay.

Per cent

calcined

clay.

Per cent
feldspar.

Per cent
whiting.

Per cent
flint.

35

55

20.5

15

230

43

55

5

20.5

I 5

18.O

53

50

20

19.O

I 5

9 5

As noted, these bodies contain 55 per cent, 60 per cent and 70
>er cent of clay, respectively. The burning curves for Nos. 35
ind 43 are very much alike. Body No. 53 contains less fluxing

614

JOURNAL OF THE

materia] and hence has a somewhat smaller shrinkage but other-
wise it is similar to the others.

Referring again to Fig. 2, the burning curves for body No. 19,
containing 45 per cent clay and 22 per cent feldspar, is also sim-
ilar to that of these bodies.

It is interesting to note the similarity in the curves for the
bodies represented in Fig. 4, which have variations in clay content

/200 /230 /Z60 /2S0 /320 /JS0 /380 /4/0 /440 /47<?

T£/fP£/?/ir(/ft£ /fi-'C
Fig. 4.

from 55 to 70 per cent but the same content of flux, to the bodies
shown in Fig 2, which have the same clay content but a wide
variation in the flux content. Although body No. 53 does not
have exactly the same content of flux as Nos. 35 and 43, it is
sufficiently close for comparison and it will be noted that it is a
little more refractory than bodies Nos. 35 and 43.

Fig. 5 is an enlarged sketch of the field covered, as shown in]
Fig. 1 , and shows the maturing points of the bodies in the different
areas.

It will be noted that the bodies containing 45 and 50 per cent
clay vitrify at a slightly higher temperature than those contain-

J

AMERICAN CERAMIC SOCIETY. 615

ing 55 and 60 per cent clay. This, however, is due to the presence
of 1V2 per cent whiting in the bodies containing 55 per cent and
more of clay. The effect of the whiting is shown later.

In this connection, it is of interest to note the softening points
of mixtures of Zettlitz kaolin, quartz and feldspar as determined

65% 70% 7S% &0% 3S<& 90<£

01/7 r oo/yrr/yr

or j?/fff/?£/yr /i/xrvxzs.

Fin. 5.

by M. Simonis.1 These have been plotted on the triaxial dia-
gram, Fig. 6. The field covered in Fig. 5 is shown in the shaded
area of Fig. 6. There is very little variation in the softening
points between 45 and 70 per cent kaolin with feldspar variations
from i21/2 to 30 per cent. There is also more variation between
12V2 per cent and 30 per cent feldspar with 45 per cent clay
than with 70 per cent clay.

The Effect of Ball Clay.

As previously stated, ball clay was not used in any of the bodies
prepared in the foregoing work. On account of its general use
and its bonding strength in the green state, tests were made in
order to show its effect both in the green state and in the burning.
In Table 4 is shown the composition of several bodies, their mod-
uli of rupture, water of plasticity contents, etc. The first four
1 Sprechsaal, 1907, No. 40.

•6i6

JOURNAL OI« TIIK

nixTutfts °orav/7/?rz.
Vg-/5g£gaZg md moan

•^ ^ ni/nasRS Rrrf/r ro co/ns.

104 Z04 304 40*/, SOX 604 704 304 S04 /004

CL/ir co/trf/ir

Fig. 6.

are all-kaolin bodies and the remainder of the series contain from
o^to 1 6 per cent of Tennessee ball clay. Of the first four, two

Table 4.

Showing the effect of ageing of bodies containing no ball clay as compared
to green bodies containing ball clay.

Body
No.

Time (sh
aged.

Drying
rinkage by
volume).

Per cent
water in
terms of
dry wt.

Modulus
of rup-
ture, lbs.
per sq. in

Per cent
ball
clay.

Per

cent
total
clay.

Per
cent
feld-
spar.

26 old

8 months

19-54

28.12

267 .6

0

50

20.0

26 new

3 days

13-98

27.80

139-8

0

50

20.0

34 old1

8 months

22 .67

31 -23

278.8

0

55

18.5

34 new1

3 days

14 -75

28.70

165.O

0

55

18.5

139 new

3 days

10.25

25-81

I57-0

0

45

20.0

140

3 days

11.99

26.44

186.O

4

0

45

20.0

141

3 days

12.07

25.92

201 .5

8

0

45

20.0

142

3 days

13-60

26.24

202.0

12

0

45

20.0

143

3 days

14.02

26.32

250.O

16

0

45

20.0

1 Bodies Nos. 34, old and new, contain 1 Va per cent of whiting.

AMERICAN CERAMIC SOCIETY. 617

vere molded into test pieces three days after preparation and
luplicates of these were also molded into test pieces after ageing
or eight months — having been kept in a damp closet and in work-
ible condition for the entire time. None of the other bodies were
iged.

In both cases when the body was aged the all-kaolin bodies
howed a marked increase in strength with age. Ageing in-
Teased. the modulus of rupture of body No. 26 from 139.8 pounds
o 267.6 pounds per sq. in. and of body No. 34 from 165 to 278.8
>ounds per square inch. In both cases the aged kaolin bodies
vere stronger than body No. 143, containing 16 per cent ball clay,
he modulus of rupture of this body being 250 pounds per sq. in.
t may be of interest to compare the strength of these bodies with
he strength of commercial bodies. Several commercial china
>odies were tested and their moduli of rupture were found to
rary from 170 pounds to 220 pounds per square inch. Several
ittempts were made to model difficult shapes from the un-aged
:aolin bodies. Comparing this with the modeling with the aged
:aolin bodies the difference was remarkable.

This is not a proof that ageing will improve the working prop-
rties of all bodies, but in view of the fact that it is desirable to
nake ware as white as possible, it is well to reduce the ball clay con-
ent of bodies to as low a figure as possible. If ageing will make this
>ossible it is well worth doing. Several potters, who grind their pot-
ery bodies exceedingly fine, claim that they become unworkable if
iged and that the clay at the bottoms of the piles becomes "rub-
>ery," as does clay that contains an excess of alkaline salts. This
s quite possible as the fine grinding may reduce the particles of
eldspar to such an extent that a part of the salts will become
oluble in time and act as casting salts do. If this is true, it
hen becomes a case of sacrificing one quality for the other and
>f depending entirely upon what the potter considers the best
hing to do. On the other hand, it may be possible to neutralize
he alkaline salts by the addition of acid before grinding.

The effects of increasing the ball clay content and the length
>f time of ageing of bodies are well marked in the burning. In
?ig. 7 is shown the volume shrinkage and porosity curves for
>odies Nos. 139 to 143, inclusive. The curves show the very

6i8

JOURNAL OF THE

marked effect the ball clay has on the tendency of the body to
overburn as well as how increase of hall clay tends to close the
body up at lower temperatures as shown by the porosity curves.
Bearing in mind that body No. 139 has about the same composition
as body No. 17, it might be well to compare the curves of Figs. 2
and 7. The two sets of curves have quite different character-
istics. In Fig. 2 is shown the effect of increasing amounts ol
feldspar. The increase not only causes the body to overburn at

o* a/rrwzan jm/t.

#00 /Z30 /2£0 /2$0 /320 /3S0 /JS0 /4/0

Fig. 7.

/440 /4T0

a lower temperature but causes it to mature at a lower tempera-
ture— the firing range remaining about the same. However,
the addition of ball clay, although it matures the body at a lower
temperature (compare porosity curves, Fig. 7), also causes it to
overburn much earlier in proportion. Up to a certain point, the
addition of ball clay will permit of the reduction of the feldspar
content.

When these results are compared with those shown in Figs. 8
and 9, the effect of ageing upon the burning behavior of bodies

AMERICAN CERAMIC SOCIETY.

619

1

I

1

!

-V0

W//1KR6C.

OLD

Br mm wwm 'wwjv

22674,
747S%

UflTfR

ftmitf

J/.2JK
28.70%

2783
/6S0

/200 /2J0 /260 7290 7320 A3S0 /360 74/0 7440 7470

Fir;. 8.

JO

/200 72J0 7260 /2f0 7J20 7JJ0 7Ja0 7470

Tf/ypf/r/jn/fff //1-'C.

Fig. 9.

7440 /4r0

620 JOURNAL OF THE

Nos. 26 and 34 is apparent. The aged bodies close up somewhat

earlier and their shrinkage is less. The curves would also indi-
cate that the ware burns much more uniformly. It lias been seen
that the difference in the transverse strength of the fresh and aged
bodies is greater than that of the body containing zero and that
containing 16 per cent ball clay. At the same time the burning
range of the fresh and aged bodies is, if anything, in favor of the
aged or tough body. On the other hand, in the case of ball clay
bodies, the increase of ball clay has a marked effect in the short-
ening of the firing range. This, then, is still another point in
favor of the use of bodies having a minimum content of ball clay
but which are improved by ageing. All of the above observations,
especially that the aged bodies shrink less than the fresh ones,
show that the ageing has a profound influence upon the colloidal
portion of the body. If ball clay must be used, as little as pos-
sible should be employed.

The Effect of the Addition of Small Amounts of Magnesium,
Calcium and Barium Carbonates.

Four sets of bodies were made up with the following body as-
a base:

6 per cent ball clay
39 per cent kaolin
18 per cent feldspar
37 per cent flint

This body was selected as it contains about as little ball clay
as is ordinarily used commercially and a fairly low content of
feldspar. Small amounts of flux were introduced for the pur-
pose of securing bodies having lower maturing temperatures. In
each case the feldspar was replaced by equal amounts of another
flux. The amounts added are given in Table 5.

Series 1.

Ta
Series 2.

BLE 5-

Series 3.

Series 4.

Body Per cent
No. flux

Body
No.

Per cent
flux.

Body
No.

Per cent
flux.

Body
No.

Per cent
flux.

164 0.2 MgO

168

0.2 CaO

172

0.2 MgOCaO

176

0.2 BaO

165 0.4

169

0.4

173

0.4

177

0.4

166 0.7

170

0.7

174

0.7

178

0.7

167 I.O

171

I.O

175

I.O

179

1.0

AMERICAN CERAMIC SOCIETY.

621

The alkaline earths were introduced as the carbonates. In the
lird series, equal equivalents of CaO and MgO were added as
irbonates.

The curves of Figs. 10, 11 and 12 show very well the effect of
le fluxes upon the burning range. The curves for bodies Nos.
j 2 to 175 are not shown as they are similar to those of
Ddies Nos. 164 to 167. The principal points to be noted are that
aO shows no noticeable action other than a slight tightening
r the body while CaO and MgO produce a very decided effect,

#00 /230 /2S0 /2S0 /320 /3Sa /300 /4/0 /440 /4Z0

Fig. 10.

ie MgO, however, being the more active. It is also very notice-
ble that the CaO does not hasten the maturing of the body but
luses overburning at a lower temperature. Its presence also
lortens the firing range considerably- On the other hand, MgO
luses the body to mature much earlier and brings about the
verfired state at a lower temperature — the former effect being
ie more prominent. The mixture of the two oxides behaves
lmost exactly like MgO by itself.
These facts would point to the conclusion that the addition of

0J2

JOURNAL OF Till.

V00 /230 /?S0 /2S0 S3Z0 &*0 /300 '4"? /**# W

FlG. I I

/e00 /230 /260 /Z00 /320 /3S0 /330 /4/0 /440 /4r0

Fig. 12.

AMERICAN CERAMIC SOCIETY. 623

. small quantity of magnesite or dolomite would be very beneficial
tut that the addition of whiting would tend to hasten overburning
ven though it would assist vitrification of the body. BaO seems
o render no useful service. It is also noted that volume shrink-
ge is greater with the addition of the MgO and MgOCaO than
rith the addition of CaO or BaO. Those bodies from No. 164
0 179 were made up into ware by several methods.

The body containing no flux, beyond the 18 per cent feldspar,
howed no translucency when fired at cone 1 1 , while the bodies

0 which were added the alkaline earths (with the exception of
he barium carbonate) evidenced very marked translucency in-
reasing with the amount added, although there was not a great
ifferenee between the effects of the 0.7 per cent and 1.0
er cent additions. The BaO produces no translucency. With
lcrease in translucency there was noted a marked change in color
rom yellowish to bluish white, particularly in ware that was
urned under reducing conditions, although this was also observed

1 the ware fired under oxidizing conditions.

The addition of small amounts of magnesite or dolomite is
eneficial in many ways, particularly to those bodies which should
lature at from cone 10 to 11.

Special Bodies.

In view of the fact that hard-fire porcelain bodies develop
lore or less well-defined crystalline sillimanite, it was decided to
splace all of the quartz in some bodies by artificial sillimanite —

mineral having a low coefficient of expansion and not subject
d crystalline transformations as is the quartz. The silliman-
:e calcine was prepared in several ways but the most satisfactory
rocedure was to calcine a thoroughly blended mixture of 258
arts of kaolin and 102 parts of anhydrous alumina. These con-
tituents were ground dry in a ball mill with 2 per cent by weight
f powdered boric acid, mixed with water, pugged into a stiff
lass, and then molded into small lumps about an inch or two in
iameter. These were calcined to cone 20. The boric acid accel-
rated the reaction and apparently did no harm — probably vol-
tilizing at a fairly low temperature. The calcined material was
round dry to pass a 100 mesh sieve and was then introduced
ito the bodv in the same manner as the other materials.

624

JOURNAL OF THE

A petrographic examination of the calcine showed the pre
ence of about k> per cent of large sillimanite crystals (o.oi to

0.03 nun. long), 40 to 50 per cent of the ground mass being weakly
bi-refringing and probably composed of very minute sillimanite
crystals too small for accurate measurement. There were also
traces of what appeared to be a minute amount of uneombined
Al,(),.

A body having the following composition was prepared:
Body No. 50.

Sillimanite calcine 22 .5 per cent

Kaolin 51.0 per cent

Feldspar 25.0 per cent

Whiting 1.5 per cent

The burning range is shown by the curve in Fig. 13 and requires
no explanation. Its long temperature range should be noted.
The sillimanite having been introduced in the place of flint, the
body is free from the strains due to volume changes caused by the

B

SPEC/ A L B0DYW50.

cof/r/i/ff5 rto/Trffj/uc/].

O00 /Z30 /2£0 /2£>0 020 /JS0 /3S0 /■?/<? /440 /•/"<?

Fig. 13.

1

\

1

AMERICAN CERAMIC SOCIETY. 625

xansformation of quartz. vSillimanite has proven useful in the
Dreparation of bodies of different compositions — particularly
Adhere a low coefficient of expansion is desired. It is to be hoped
:hat natural deposits of this mineral will be found in quantity.
ft has also been found that the sillimanite calcine may be in-
corporated with the flux used in the body — which makes it pos-
sible to reduce the calcination temperature to as low as the
softening temperature of cone 14. A natural deposit, however,
tvould be ideal for use providing the content of iron oxide were
sufficiently low.

Cone 11 Bodies.

A number of the bodies in the field covered are within
the range of commercial bodies fired from cones 10 to 12, even
though the clay content be as high as 60 per cent. All things
considered, the range of body composition for ware fired to from
:ones 10 to 12 (such as hotel china) would lie between the follow-
ing limits :

Flint 30 to 40 per cent

Feldspar 14 to 18 per cent

Alkaline earth fluxes 4.0 to 0.5 per cent

Ball clay 3 to 8 per cent

China clay 45 to 25 per cent

Florida clay 10 to 6 per cent

It is evident, of course, that it would not be permissible to use a
high feldspar content together with a high alkaline earth content.
Dolomite is found in nature sufficiently pure for use in pottery
bodies and can be secured at a very low cost. The lime, mag-
nesia, or dolomite fluxes are of considerable aid in securing a
whiter color.

Florida kaolin may be used to a certain extent as a substitute
for ball clay, but should not be depended upon entirely in the
manufacture of thin or medium-weight ware.

The limits given for ball clay are lower than some potters are
probably using but they are as high as is necessary under working
conditions where reasonable care is used. This statement is
made after careful consideration and actual knowledge of the
working qualities of such bodies.

626 JOURNAL OF THE

Color.

In many wares, and particularly in table wares, color is one ol
the most important factors connected with their sale and urn si
receive serious consideration. Sometimes other qualities must be
sacrificed in order that the color may be as nearly white as pos-
sible. This is to be regretted but must be contended with as long
as the trade demands it. vSome of the most beautiful ware made
in America is cream colored and is finding a large market. There
is nothing more pleasing than cream-colored ware and it is a
mystery as to just why china should be white.

There are three ways of securing white porcelain: first, the use
of very pure materials; second, burning under reducing condi-
tions; and third, masking the yellow color with a blue stain. The
first method is ideal but practically out of the question. How-
ever, the purity of all materials must be considered and the
quality maintained at as high a standard as the grade of ware
will warrant.

As previously stated, the addition of small amounts of alkaline
earth fluxes, such as dolomite, magnesite or whiting, improves the
color regardless of any other treatment the body may receive.

The second method is largely used in Europe and when properly
carried out results in a purer white than is produced by a stain.
The ferric oxide is reduced to the lower oxides which, owing to
their dark color and fineness of grain, produce a whiter effect.
Care must be taken to start the reduction at a sufficiently low
temperature so that the reducing gases may penetrate the
structure completely and thus bring about the desired effect.
The bisque burn is carried to a low temperature and the glost
burn brought to maturity in the following way: The burn is
divided into three parts, viz., the heating-up period, the reducing
period, and the finishing period. Oxidizing conditions are main-
tained during the heating-up period — which extends from the
beginning of the burn to the softening point of cone 09.

The reducing period extends from cone 09 to cone 8 — the latter
point depending somewhat upon the constitution of the body and
of the glaze. During the reducing period, carbon monoxide and
hydrogen should be present in the kiln gases in quantities from
3 to 5 per cent. Too heavy reduction should be avoided. If

AMERICAN CERAMIC SOCIETY. 627

duction is too strong when the glaze begins to soften and the
)dy begins to tighten up, the ware may be spoiled by the deposi-
Dn of carbon in the body or softened glaze.1
The third procedure, the use of cobalt stain, represents the
miliar effect of the bluing used in washing clothes. It is evi-
;nt that the use of one color in order to mask another cannot be
satisfactory, as far as quality of shade is concerned, as the total
imination of the yellow color. However, for ordinary purposes,
ie difficulty of securing proper reduction warrants the use of a
ain. The cobalt is preferably introduced as the precipitated
tsic carbonate (obtained by the addition of sodium carbonate to
solution of cobalt salt), followed by one or two washings by
:cantation. The precipitated carbonate must then be ground
to the body in a ball mill. It may also be introduced into a
)dy in a blunger but this practice is not as satisfactory.
In conclusion the writers wish to express their appreciation to
r. A. V. Bleininger for advice in outlining the work and to the
rce of the Clay Products Division of the Bureau of Standards
r a great deal of assistance in carrying on the experimental work.

1 For complete description see G. H. Brown, "Burning of Porcelain,"
oceedings New Jersey Clay Workers' Association, 19 14.

THE USE OF CAR TUNNEL KILNS FOR BRICK AND
OTHER PRODUCTS OF CRUDE CLAYS.

By Ki.r.is I.ovKjoy, KM, ColumbllB, O.

There are a number of problems in ear tunnel kilns which mus
be solved in order to adapt the kiln to eommon clay products.

A major problem is to distribute the heat uniformly among
the ware. The ware is in relatively small units but each unit
comes in close contact with the furnace.

In large down-draft kilns the ware placed around the bags
where the flame strikes it is largely rejected or at best it goes into
second quality or culls. There would exist the car tunnel kiln
problem in down-draft kilns if the kilns were built four to six
feet wide with the bags removed and the ware set in small units
in front of the furnaces. We would have then about one thousand
bricks with a furnace on each side. How many of the bricks
would be fire-flashed and overburned? Suppose we are burning
face bricks, what percentage of product with uniform or satis-
factory color would be obtained? Color is a very sensitive test
of combustion conditions. Paving bricks are also very sensitive
to heat conditions. Many factories cannot produce No. i qual-
ity bricks around the bags and over the top of the kiln, but in
a down-draft kiln the percentage of these second quality bricke
is relatively small. What would be the result if fifty or more
per cent of the product is exposed to the flame?

Pottery, in the manufacture of which the car tunnel kiln i;
making rapid progress, is placed in saggers and the walls thereo
interpose between the flame and the ware. Moreover, the siz<
and shape of the saggers are such as to provide large spaces foi
the circulation of the air and gases. In some operations it i:
customary to set bisque on top and glost at the bottom, thu
taking advantage of difference in temperature, and such differ
ence in temperature is not serious.

Bricks, however, are set in close checker work and it is difficul
to distribute the heat uniformly through the mass.

AMERICAN CERAMIC SOCIETY. 629

We cannot set bricks vertically around the tops of the bags in
)vvn-draft kilns because the greater shrinkage on the flame face
ill cause the face to draw over and fall into the bags, followed
i a second and a third face as they become exposed to the flames.
re must, therefore, rack back the bricks on cars as we do around
In bags, and if we do this where are we to get any high-grade
ce bricks which must be perfectly faced in the setting? Some
)llow tiles made from high shrinking and low burning range
ays are badly warped and misshapen where the flame strikes
tern and the loss of ware would be serious if the flame struck
le bulk of the mass.

Any cull brick pile will contain bricks much smaller on one end,
le to the fact that the small end was exposed to flame, as in the
ches of up-draft kilns. Under such conditions there is a marked
fference in color, size, and hardness in eight inches in length.
rould the conditions in a car tunnel kiln be equivalent to those

the arches of an up-draft kiln ?

In three designs of car tunnel kilns the combustion takes place

a box or flue on each side of the kiln and the heat conducted
irough the walls of the flues is carried to the ware by circulation

air around the combustion flues thence over and down among
te ware. In two kilns the operation is regenerative — the air
ter passing among the ware re-circulates around the combus-
an flues. In two other designs the operation is that of a down-
'aft kiln in which the combustion gases rise through a bag and
;scend among the ware and are drawn through a perforated
x>r into the next compartment and continuing until finally
'awn from the kiln. In the other designs the cars of ware sim-
y pass a flaming furnace sometimes with an intervening flash
• perforated wall. There are then three types of car tunnel
Ins — direct heating, indirect heating in the tunnel, and com-
irtment operation.

One direct-heating type sets a combustion arch transversely of
le ear in the bottom of the ware comparable with the arch in
1 up-draft kiln and this arch, being built of fire bricks and re-
lined as a permanent part of the ear furniture, should give better
:sults than those from the arch of an up-draft kiln. The in-
irect-heating types would seemingly better solve the flame

630 J< >1 RNAL OF THE

difficulty, although the compartment type appears advantageoui
and may prove very efficient when it is successfully developed and
adapted. Unfortunately, one of the compartment type was not
successful in an initial installation, but one must not conclude that
the type is not promising because of this single failure.

There have been many failures of the direct-fire tunnel types
in the past, yet new designs are constantly coming and provini
successful in some lines of ware.

Manufacturers of common wares will have to be shown that
simply pushing cars of ware through a fiery furnace will give satis-
factory results: — namely, freedom from fire-flashing; uniformity
in color, size, and hardness; no warping and no fire-cracking.

Assuming that the flame difficulty may be overcome in some
types of kiln, there remain the racking back, complex bonding
to insure the stability of the mass of ware and to provide circu-
lating flues.

Such setting is serious in face brick manufacture and may b«
objectionable in other common wares. Drain tile, for instance
does not lend itself readily to complex setting, and it is not desir
able to have transverse or longitudinal flues in the ware.

The manufacture of common wares is regarded by many cer
amists as a job requiring little intelligence, but as a matter of fac
the common ware manufacturer has to overcome the most diffi
cult problems. He is dealing with a single raw clay with all it
inherent difficulties which must be overcome in a large way ii
every step of the process of manufacture, while the manufacture
of other wares is dealing with a synthetic body and is incompeten
if he cannot improve on natural clays.

One of the problems in a car tunnel kiln is to protect the car
from the heat. Sand seals on the sides and tongued and groove*
end-locks in the car frame, or furniture, shut off the cars from th
heat tunnels fairly well but not perfectly and they do not ovei
come conduction. Two designs do not even use sand seals ann
the car simply projects into a recess in the side wall, at best offe:
ing but slight resistance to the upward passage of air. In pra<
tically all designs there is some method of cooling the und(
tunnel and thus protecting the cars. One cools the cars by a
pipes along the tunnel walls adjacent to the car structure. Th

AMERICAN CERAMIC SOCIETY. 631

in indirect way of cooling by means of which there is theoreti-
ly no air circulation in the tunnel under the car floor and the
[y air which could get into the upper tunnel from the car space
uld be that from leakage. Another design has a duct under
! cars which may be closed and the only air admission is that
m leakage. Dampers in the bottom of this duct may be
ined and any amount of air allowed to enter. The amount
air which is allowed to enter is that required to keep the cars
)1, and this air circulates around the car structure and rises

0 the heat tunnel, there being no sand seal. A third design
roduces the air under pressure through pipes and jets the air
tinst the car structure at intervals in the hot zones. Again we
/e direct provision to draw the air through the car duct and
iver it under the furnace grates or to the gas burners, usually
ree circulation under the cars but in one instance through a
:et metal casing, thus cooling the cars indirectly.

t matters not how it is accomplished, the cars must be cooled

1 there is more or less air finding its way up around the cars
I enveloping the ware.

>t us consider a down-draft kiln in this light. Suppose we
:ulate cold air under the kiln floor and allow it to seep, or
jly flow, into the kiln on the inside of the wall at the floor
el. Every down-draft kiln burner knows, or should know,
,t chilling the kiln bottom results in serious loss in the bottom
re. A prominent company several years ago put flues under
: kiln floor through which air was to be taken from the out-
e, under the kiln floor, heated by conduction from the waste
abustion gases, then back to the furnace for primary combus-
ti. The purpose was to heat the air for combustion which it
omplished, but at the same time cooled the kiln bottom to
extent of serious effect on the ware and the scheme had to be
indoned, as have several other similar schemes.
Dampness under the kiln has the same cooling effect on the
1 bottom in consequence of evaporation and many manufac-
ers have experienced trouble and losses on this account. Com-
n wares cannot be burnt satisfactorily in either down-draft
continuous kilns if the kiln floor is chilled and, particularly
h some short-burning range clays, the effect would be to quickly

63 ' J< >URNAX OF THE

put a profitable business into the hands of a receiver with an
ultimate sale of the plant for junk.

The car tops arc cooled and the best of insulation is not the
equal of a blanket of high temperature combustion gases under
the kiln floor. In the combustion zone we have in the direct-fire
kilns combustion gases rising in the space between the ware and
the flash wall of the furnace. We also have a current of cold air
from under the cars creeping up the face of the ware, taking heat
from the bottom ware and carrying it to the top and finally com-
mingling and combining with the furnace gases in secondary
combustion, thus still further increasing the temperature on top.

There are many clays, particularly limey clays, and some vit-
trified products, which cannot be profitably burned if the varia-
tion in kiln temperature exceeds 6o° C, — three cones. This can
be done commercially in down-draft kilns. We say "commer-
cially" with a purpose. There can be obtained the same cone
temperature in top and bottom of a down-draft kiln but at such
fuel expense as to rule it out as a commercial operation. What
can be done commercially in a car tunnel kiln? Certain reports
indicate the same temperature throughout the mass of ware.
This sounds encouraging indeed, but is it true? The manufac-
turers of crude clay wares wish to know and are greatly interested.
There are also reports that common ware on the face of the cars
is vitrified, or at least extremely hard, while that in the bottom
center is barely within the limit of salable building bricks. This
does not check with reports previously mentioned. Some kilns
are burning all biscuit and others all glost, but what variations in
temperatures will these wares stand? One kiln, at least, burm
bisque on top and glost in the bottom to take advantage of the
difference in temperature.

L. E. Barringer1 states that, "the kiln at Schenectady is used I]
firing porcelain insulators to a temperature of 1300 to 1400]
C." A variation of 100 degrees in a four-foot column of commoi
ware in many instances is prohibitive.

It is frequently necessary to get within 60 degrees in a ten-foo
column of ware, and it is desirable to do better. Even this varia
tion, in a four-foot column, would give too much difference ii
1 Trans. Am. Ceram. Soc, 18, 118 (1916).

AMERICAN CERAMIC SOCIETY. 633

)lor, size, and degree of vitrification to be satisfactory. In
Feet, the product would be practically all "bag" bricks, since we
mnot encase our common wares in saggers.

Common wares are often quite wet when set in the kilns. They
lould not be, of course, but they are, and in consequence require
ireful water-smoking, which is particularly true of dry-pressed
ricks requiring from three to eight days of slow water-smoking,
''ill the car tunnel kiln be adapted to this product ?

Many common clays are high in carbon and iron sulphide, as
tr example the black lignitic clays in New Jersey and the black
hio shale. The oxidation of these is a serious problem. How
ill it be worked out in the car tunnel kiln?

One might assume that we are opposed to the car tunnel kiln,
it, quite the contrary, we are very much interested in its success
id when we are shown that it will advance the manufacture of
jmmon clay wares, we will boost it to the limit.

The car tunnel kiln is a century old and more, and there are
lany wrecks strewn along the wayside. In the present revival,
le kiln has come to stay, but there are a number of problems
hich must be solved to adapt it to the successful production of
)mmon clay wares.

We have not a complete list of the car tunnel kilns in this coun-
y, but there are over forty under construction, in operation, or
aandoned.

So far as indicated by available records there are fifteen under
mstruction and in operation on white-ware, porcelain, etc.;
lere are five burning refractories; two on building bricks; seven
Dandoned as failures. There are also several in other industries,
-burning carbons, annealing, etc., about which we have no in-
)rmation.

There are more than a dozen kilns in this country devised by
ifferent inventors, from which the clay workers may choose,
f these, at least ten have been tried out, or are under construe-
on. Two types have the credit of twelve or more kilns each in
lis country. This shows the marked interest in the car tunnel
iln.

The kiln is rapidly being given the opportunity to prove its
lerits, but much remains to be done to adapt it to the several

634 JOURNAL OF THE

distinct lines of ware before il can take its place alongside of, or
displace, the other types of continuous kilns.

COMMUNICATED DISCUSSION.

L. E. Barringer: The author has clearly pointed out some
of the difficulties which might be anticipated in attempting the
use of continuous tunnel kilns in the burning of products made
from crude clays, as structural and paving brick, drain tile, etc.

The first criticism which is made is that of possible lack of a
sufficiently uniform temperature in the continuous kiln and the
difficulties of properly distributing the heat through the ware as
the latter passes through the combustion zone. It is pointed out
that the nearness of crude clay wares to the throats of the fire
boxes in all types of kilns leads to fire-flashing, overburning and
shrinkage differentials, and it is the author's opinion that a very
much larger percentage of the product would be subjected to
overfiring, through nearness to the fire boxes, than in the large
periodic (and principally down-draft) kilns commonly employed
for the firing of the crude clay wares.

Exposure to the direct flame of the furnace is of course produc-
tive of high losses and the author's question as to what would
be the result if 50 per cent or more of the product should be
exposed to the flame, might be answered by saying that the
result would probably be a loss of high-grade ware to the extent
of 50 per cent or more. But there should not be such a high
percentage of ware exposed to excessive temperature and I be-
lieve that the distribution of temperature can be controlled within
closer limits in a properly designed tunnel kiln than in a large
periodic down-draft kiln.

Unfortunately, my statements1 of temperature distribution in
connection with the tunnel kiln in use in the Porcelain Depart
ment of the General Electric Company were perhaps not made
sufficiently clear. In the paper in question it was stated thai
"the kiln at Schenectady is used in firing porcelain insulators to i I
temperature of 1300 to 14000 C" which was intended to mear
that the maturing temperature of porcelain was always betweer
these limits, that is, some porcelains were fired at a little highe:
1 Trans. Am. Ceram. Sot., 18, 118 (1916).

AMERICAN CERAMIC SOCIETY. 635

emperature than others. A little later in the same paragraph,
lowever, the statement is made "a uniformity of distribution of
emperature is secured of well within one cone." This means
hat whether the porcelain is fired at 13400 C or 13800 C, the actual
lifferential at the proper maturing temperature can be reduced
o within 20 ° C. For instance, if the maturing temperature is
340 ° C, the temperature in all parts of the firing zone of the
ontinuous kiln can be maintained within the limits 1 330 to
350° C.

In the same paper the writer made the statement that for brick
md crude clay wares, the necessity of uniform temperature is
iot as great as in the higher-grade clay wares, since matters of
(orosity, exact size and uniform color are not as important. I
hink the author of the paper under discussion confirms this in
tating that there are many common clay products which cannot
>e profitably burned "if the variation in kiln temperature exceeds
io° C (3 cones)." This is decidedly more than the one cone or
o° C which is required in the firing of such ware as electrical
>orcelain and which is maintained in the tunnel kiln being oper-
ted at Schenectady.

There would seem to be no fundamental reason why the crude
lay wares cannot properly be protected from fire-flashing and
»ver-burning in a continuous tunnel kiln if the fire boxes and fire
one are properly constructed. Inasmuch as the temperatures
equired in firing brick and similar products are not as high as
hose necessary for the firing of porcelain, the fire boxes can be
ilaced at a greater distance from the tunnel of the continuous
:iln at the firing zone without excessive loss of heat.

A second point that the author brings out is the uncertainty of
>eing able to regulate the condition of the fire gases and to bring
.bout oxidizing or reducing conditions as required. This, how-
ver, can be done, although it must be admitted that a contin-
tous kiln from its very nature cannot be as quickly manipulated
11 this respect as a periodic kiln and the changes from a reducing
ire to an oxidizing fire or vice versa require longer periods. Such
ontrol can be obtained, however, either in the open tunnel kiln,
is the Didier-March, or in the muffle kiln, as the Dressier.

In discussing the matter of protecting the cars by sand seals

636 AMERICAN CERAMIC SOCIETY.

and otherwise, and in considering the possibility of air leakage

which should tend to cool the floor or bottom of the kiln, tin
author fears that such cold air leakage will materially affect the
temperature of the ware at the bottom of the kiln. In this con-
nection, however, it is to be pointed out that the very point where
such air leakage might take place is at the point of maximum
temperature, that is, between the face of the ware and the ports
from the fire boxes, so that slight air leakage might possibly do
more good than harm. Of course, if such leakage is excessive,
it may be injurious, both to the under-structure of the car and to
the ware upon the car. If the writer understands the paper cor-
rectly, however, the direction of cold air leakage is considered as
upward, whereas designers of kilns with sand seals undoubtedly
assume that the leakage will be downward into the subway be-
neath the cars and the sand seal is particularly necessary to pre-
vent the under parts of the car becoming so overheated that they
are not as readily operated, to say nothing of the rapid deteriora-
tion of the iron and steel under such conditions.

It seems to the writer that of the difficulties which have been
pointed out by Mr. Lovejoy, the most troublesome to meet suc-
cessfully would be (a) the proper "water-smoking" of crude clay
wares, (b) the handling of highly carbonaceous, calcareous and
pyritic clays, and (c) the control of kiln gases, within sufficiently
short time periods.

The designers and builders of continuous car tunnel kilns will
probably desire to discuss the proper construction and method of
handling such kilns when the burning of crude clay ware is in-
volved and to answer some of the pertinent inquiries which Mr.
Lovejoy has made.

General Electric Co.,
Schenectady, N. Y.

HE PRESENCE OF IRON IN THE FURNACE ATMOS-
PHERE AS A SOURCE OF COLOR IN THE
MANUFACTURE OF OPTICAL GLASS.

By Edward W. Washburn, Urbana, III.

It is well known that the presence of very small quantities of
on in glass imparts to it a greenish color. In the common varic-
es of glass, this color is usually eliminated by employing a "de-
jlorizer" which produces the complimentary color. However,
lis procedure is not possible in the case of optical glass owing to
le decreased transmission of light which accompanies it. For
us reason each material employed in making up the batch for
roducing optical glasses must have an extremely high degree
I purity as regards its content of iron.

The coloring effect of iron seems to be especially pronounced
i the case of the flint glasses, and the glasses of this type hitherto
lanufactured in this country for optical purposes all have a very
/ident greenish tinge and this in spite of every precaution to
nploy only the purest materials. The writer's attention was
rst directed to this situation by the results obtained with some
cperimental melts of a heavy flint glass (nD = i . 64) obtained
l an electric furnace. These melts were entirely free from any
ace of color, although the same batch when melted in the same
rpe of pot in a gas-fired furnace invariably gave a product hav-
ig a decided greenish color.

The pots in both cases were kept carefully covered during the
lelting and the other conditions in the two experiments were
ich that the only possible source for the color obtained with
le gas-fired furnace seemed to be the presence of iron in some
>rm in the atmosphere of the furnace. An opportunity to further
:st this hypothesis was afforded the writer by Mr. A. V. Blein-
iger, of the Bureau of Standards. The tests were made on one
f the large pot furnaces of the Bureau's optical glass factory
t Pittsburgh and were conducted in the following manner:

A five-foot fused silica tube, open at both ends, was inserted

638 JOURNAL OF THE

through the wall of the furnace so that the cud of the tube was
immediately above the pot of melted glass. The temperature
of the furnace was approximately at the maximum point attained
during the manufacturing operation, and was just immediately

preceding the period at which stirring was begun. The outer
end of the tube was connected to a glass tube ending below tin-
surface of a dilute solution of sulphuric acid contained in a wash-
bottle. Suction was applied to the wash-bottle so as to draw
a stream of the furnace gas through the sulphuric acid. The
experiment was continued for two hours, at the end of which time
the silica tube was removed and washed thoroughly on the in-
side with some of the sulphuric acid from the wash-bottle. The
acid was then tested for iron in the usual way with a solution of
ammonium sulphocyanate, and a distinct positive test was ob-
tained.

This experiment clearly demonstrated the presence of iron in
the furnace atmosphere and rendered it extremely probable
that this constituted a significant source of the color in the glass.
This was further confirmed by the fact that, previous to stirring,
a sample of the glass removed from the pot had a much better
color than was the case after the pot had been stirred for a num-
ber of hours. Evidently, the iron taken up from the furnace
atmosphere was mixed throughout the glass by the process of
stirring.

The iron in the atmosphere was doubtless present in the form
of ferric oxide, but whether the oxide was present in the form of a
fine dust or in the form of vapor was not definitely determined,
although it seems reasonable to believe that at the temperatures
attained in the furnace, ferric oxide might have a sufficiently
high vapor pressure to produce the observed results. However,
it is hoped to settle this question definitely by attempting to di-
rectly measure the vapor pressure of ferric oxide at high tempera-
tures.

Two obvious sources for the iron in the furnace atmosphere
were: (1) The iron burners which extended far enough into
the furnaces so that the ends were covered with a coating of the
oxide; and (2) the iron content of the fire-brick lining of the fur-
nace. In order to remove these two sources of contamination,

AMERICAN CERAMIC SOCIETY. 639

ie iron burners were replaced by burners made of clay and the
terior walls of the furnace were covered with a coating of kaolin.
s a result of these changes in the furnace, Mr. A. E. Williams,
iperintendent of the Pittsburgh Optical Glass Plant, reports
at a distinct improvement in the color of the heavy flint glasses
ts resulted.

It was noted that the layer of white kaolin which was placed
1 the interior walls of the furnace gradually acquired a red color,
ther by the absorption of iron from the brick behind it or by
ie taking up of iron from the furnace atmosphere.
The results of the above experiments demonstrate very clearly
tat in the construction of furnaces for the manufacture of optical
ass it is highly desirable to exclude all materials which will give
) iron to the atmosphere of the furnace. The interior of the
mace should, if possible, be lined with a course of kaolin bricks,
id if it is necessary to employ any object constructed of metallic
Dn for any purpose in the interior of the furnace, water cooling
lould be employed to prevent volatilization of the iron.

Department of Ceramic Engineering,
University of Illinois.

THE CONTROL OF THE LUSTER OF ENAMELS."

By iiomkk f. Stalky, Washington, D. C.

Crystallization.

When melted, glasses are true solutions, corresponding in
all respects to solutions of salts and water at ordinary tempera-
tures. When these molten solutions are cooled, we would expect
them to turn gradually to a mass of crystals, just as an aqueous
solution of salt does if the temperature to which it is cooled is
sufficiently low. The only difference is that the glassy solutions
should crystallize at higher temperatures. The temperatures at
which glasses would become completely crystallized, if sufficient
time were allowed, vary from 550 to noo° C, yet in practice
they are cooled and exist permanently at atmospheric temperatures
without any crystallization taking place. In this condition they
are still true solutions — differing from aqueous solutions only
in being so viscous that they are rigid. All bright finish enamels,
glasses and glazes may be regarded then as undercooled liquids.
The tumblers from which we drink are in a sense just as truly
liquids as the water they contain.

Of course, if the constituents of a glass crystallize, the surface
becomes mat instead of bright. Since bright finish enamels
are the type desired for enameling metals, the enamel maker must
so compound his enamels that they will not crystallize when
subjected to the normal working and cooling conditions of enamel-
ing shops. Fortunately, the sudden cooling to which enameled
ware is subjected is not favorable to crystallization. On the other
hand, the working temperature is so low that the glass is easily
saturated by the compounds formed in the molten magma by
some of the oxides used. Just as a hot, saturated, aqueous
salt solution tends to become supersaturated and to deposit
crystals of the salt on cooling, just so does a hot, saturated glass
or enamel solution tend to deposit crystals when it is cooled.

Effect of Viscosity.

The most simple method of overcoming the tendency to crystal-
1 By permission of the Director, Bureau of Standards.

AMERICAN CERAMIC SOCIETY. 641

ce is to make the enamel viscous at the temperatures at which
ystallization takes place. In even a moderately viscous liquid,

takes considerable time for crystallization to start and the rate
' growth is slow. Since enameled iron ware is cooled so rapidly,
ifficient time is not given in the cooling process for an appreciable
irmation of crystals in a viscous enamel, unless the tendency
r deposition of crystals is very great.

Silica alone melts to a very viscous glass, and, in general, the
Idition of silica to any melt increases the viscosity. Silicates
gh in silica also melt to viscous glasses. If silica (flint), sand or
ldspar (a high-silica compound) is added to an enamel, the glass

rendered more viscous at all temperatures, and the tendency
> become dull from crystallization is lessened. Since the work-
g temperature of the enamel is raised by this procedure, this
ethod has but slight application in controlling the viscosity
' enamels that must mature at temperatures rather rigidly
ced by shop practice and working conditions.

Boric oxide alone forms a very viscous glass at low tempera-
ires; but, as shown by the writer,1 the viscosity decreases rapidly
5 the temperature rises. The glass was found to be three times
; viscous at 750 ° C as it was at iooo0 C. The first temperature
Dproximates that at which crystallization takes place in enamels,
id the second is near the maturing temperature of most enamels
>r cast iron. The borate glasses of any given base resemble
licate glasses in behavior and become more viscous as the per-
:ntage of boric oxide increases. The following results were
Dtained for barium oxide — boric oxide glasses:

Per cent
B2O3.

Per cent
BaO.

Relative
same

viscosity at the
temperature.

20

80

3 -OS

25

75

3.18

30

70

3-35

35

65

3 -52

40

60

3 -70

45

55

4 .02

It is noteworthy that the presence of compounds and eutectic
impositions produced no noticeable effect on the viscosity curves.2

1 "The Viscosity of Molten Glasses," Original Communications, Eighth
itemational Congress of Applied Chemistry, 5, p. 127..

2 Cf. Trans. Am. Ceram. Soc, 13, 675 (191 1).

642 JOURNAL OF THE

The viscosity of borate glasses decreases with rise of tempera]
ture in the same general way as thai of fused boric oxide, but

not as rapidly. The effect of increasing the boric oxide in cast
iron enamels is to increase the viscosity at the temperatures at
which crystallization is liable to take place, and thus to decrease
the tendency to matness. Its effect differs from that of silica
in that the maturing temperature is not raised but, in the case
of many enamels, is actually lowered.

Kaolin (china clay) has a very decided tendency to raise the
viscosity of enamels but is not much used in American although
found in some German formulas. Any material which does not
fuse or go into solution in the enamel, such as most metallic oxides
added as opacifiers, makes the enamel more viscous. Decrease
of any of the fluxing oxides, except boric oxide, makes the enamel
more viscous. Therefore, viscosity may be increased by decreas-
ing any of these. Decrease of fluorides has the same effect.
Effect of Concentrations.

With the exception of boric oxide, all of the materials in enamel
compositions known as fluxes, i. e., the oxides of lead, zinc, sod-
ium, potassium, calcium, barium, magnesium, and the various
fluorides, tend to form compounds which are easily crystallizable.
Lead oxide differs from the others in that its compounds when
melted alone do not crystallize until very low temperatures are
reached. Lead silicates do not crystallize until temperatures
of 7500 C or lower are reached,1 and lead borates have even lower
crystallization temperatures. Since normal enamel compositions
are too viscous at low temperatures to allow crystallization tc
take place, lead compounds do not crystallize from enamel glasses.
Therefore, lead oxide shares with boric oxide the distinction of
being a flux and yet not producing a mat finish when present in
the enamel composition in large amounts.

Of course, the tendency for any compound to separate in thej
crystalline form from a cooling glass is decreased by decreasing?
the amount of the compound in the glass. Therefore, the amount
of each of the fluxes, except boric oxide and lead oxide, should
be kept low. This means that as the total amount of these crystal-
forming fluxes increases a greater variety must be used.
1 Cooper, Shaw and Loomis, Am. Chem. Jour., 1909, p. 461.

AMERICAN CERAMIC SOCIETY.

643

The effect of concentration of various fluxing oxides is shown
y the series of enamels described below. As the basis of the
sries, the following batch was taken:

Potash feldspar 41 .0

Borax 24 .6

Sodium nitrate 2.7

Cryolite 12 .0

Tin oxide 8.0

Melted Percentages.
26 .5 silica
7 .5 alumina
7 .0 potassium oxide
9 .0 boric oxide
5 .0 sodium oxide
4 . 8 aluminium fluoride

7 . 2 sodium fluoride

8 .0 tin oxide

75 .0 constant
25 .0 variable

The composition of the 25 per cent variable part and the luster
f the resultant enamels is indicated in the accompanying table:
Table Showing Effect of Composition on Luster.

um-
»er.

Na20
from
soda
ash.

BaO
from
barium
carbon-
ate.

CaO
from
calcium
carbon-
ate.

ZnO
from

zinc
oxide.

PbO
from
red
lead.

B2O3
from
boric
acid.

5 OO

S-33

8.33

Luster.

I

25

mat

2

25

semi-mat

3

25

not melted

4

25

mat

5

25

bright

6

25

bright

7

12.5

12.5

semi-mat

8

12.5

12.5

mat

9

12.5

12.5

semi-mat

0

12.5

12.5

semi-mat

1

12.5

12.5

bright

2

12.5

12.5

bright

3

8.33

8.33

8.33

oily

4

8-33

8.33

8.33

bright

5

5.00

8.33

8.33

3-33

bright

3-33

bright

644 JOURNAL OF THE

Enameling Technique.

Any necessity for prolonging the enameling operation thai
involves partial cooling and reheating of the enamel after the coat
is matured furnishes favorable opportunity for the development
of crystallization and dull finish. Therefore, considerable care
must be taken to avoid all necessity for patching by having the
metal in the best possible shape for enameling before the opera-
tion commences, and by employing the most skilful workmen
to apply the enamel.

Sulphur Compounds.

It is well known that sulphur compounds in an enamel tend
to separate on the surface in feather-like crystals, causing a
diminution of luster. While this may be due to sulphur in the
enamel materials, and the presence of sulphur in these should be
avoided, it is not probable that enamel compositions made from
the grade of chemicals commonly employed contain enough
sulphur to cause the trouble. Seger has shown that a glass can
contain 2 to 4 per cent of sulphur trioxide and still be perfectly
clear.1 It is interesting to note that when Seger made a melt
from sodium carbonate, barium sulphate and silica, the glass
gall, which separated when an excess of sulphur was present,
consisted of sodium sulphate, not barium sulphate as might be
expected from the strong affinity of barium for sulphur at low
temperatures. In fact, glass gall from common glasses ordinarily
consists largely of sodium sulphate. Mellor2 records experiments
in which ordinary pottery glazes were found to dissolve compara-
tively large amounts of various sulphates, such as those of barium,
calcium, lead, sodium, and potassium, without showing any ill
effects. An experienced enamel mixer informs me that he once
had a separation of sodium sulphate in the form of glass gall
in considerable amount from an enamel batch in the melting tank.
In grinding this batch he mixed the glass gall with the good glass
and got satisfactory ware from the mixture. Mellor believes
that the feathering of pottery glazes by sulphur is due to the satura-

1 "Collected Writings of Herman A. Seger," p. 646. Published by Am.
Ceram. Soc.

2 J. W. Mellor, "Clay and Pottery Industries," Vol. 1, p. 63, published
by Chas. Griffith and Co., London.

AMERICAN CERAMIC SOCIETY. 645

on of a surface film of the glaze by sulphur in the kiln atmosphere,
his is probably the case in enamels also, for it occurs most fre-
iently when a sulphur-bearing fuel is used for heating the enamel-
g furnaces. A number of enameling plants experienced diffi-
llty from sulphur feathering when they changed from the use
: natural gas to bituminous coal. The trouble is much less
)mmon when well-ventilated muffles are used than when the air
irrents in these are sluggish. Cracks in the floor of the muffle
hich allow the products of combustion to leak in are a common
mse of sulphured ware. Open-arch or semi-muffle kilns are
rone to produce ware of poor gloss unless a strong draft is used
) as to keep the flame hugging the roof of the furnace.
Sometimes quicklime is placed around the inside of the muffle
gainst the walls, in the hope that the quicklime will absorb
le sulphur. This is generally not very effective. In the first
lace, it is only rarely that the sulphur-bearing fumes pass through
r over the quicklime unless it is placed directly over a crack or
ole. The lime does not attract the sulphur from the kiln atmos-
here as a whole. In the second place, if sulphur-bearing fumes
0 come in contact with the lime, it soon becomes saturated and
lould be replaced much more often than is the common practice,
ne per cent of volatile sulphur in a ton of coal will give 40 pounds
f sulphur dioxide or 50 pounds of sulphur trioxide, either of which
ill neutralize 35 pounds of quicklime if brought into intimate
intact with it. It sometimes pays to pile quicklime over a
imporary patch in the floor of a muffle, but even then the lime
lould be changed daily.

Index of Refraction.

Aside from the question of glasses being dulled by the presence
f crystals or sulphur compounds on their surfaces, the brilliance
f the glasses themselves will vary. The brilliance of a glass

measured by its refraction for light, the higher the index of
^fraction, the more brilliant the glass. In any one class of glasses,
le index of refraction is higher the greater the density of the
lass.1 Since lead oxide, barium oxide and zinc oxide tend to

1 H. Hovestadt, "Jena Glass," pp. 388-393, published by The Maemillan
o., New York.

(.,(• JOURNAL OF THE

produce very heavy glasses, they also produce- glasses of great
brilliance. Potassium oxide makes more brilliant glasses than
sodium oxide, and the substitution of boric oxide for siliea makes
more brilliant glasses, although the density of the glass is lowered
slightly. High-flint enamels are generally more brilliant than
those in which feldspar is the only refractory.

COMMUNICATED DISCUSSION.

E. P. Poste: We were very much interested in the remarks
under the heading "Sulphur Compounds," in which Mr. Staley
has referred to the presence of sodium sulphate on the surface of
enamels, introduced by sulphur in the raw materials entering
into the enamel mixture or by sulphur gases in contact with
the enamel during burning.

We have carefully studied this particular type of trouble under
two different sets of circumstances in connection with steel enam-
els applied by the wet process.

The first instance came to our attention while endeavoring to
work, on a semi-commercial scale, an enamel which, through the
experimental stages, had been very difficult to bring to a proper
condition for spraying. The clays which were in use in "setting
up" commercial formulas at that time did not properly suspend
the enamel in question — without the addition of an amount which
caused a definite dullness of the surface. Using less clay, it was
necessary to add such an excessive amount of magnesium sulphate
solution as a vehicle that a very definite sulphur scum was pro-
duced.

The other instance in mind came to our notice while we were
endeavoring to burn commercial formulas on an experimental
scale in a small furnace which was being fired by producer gas
under conditions such that the products of combustion came in
contact with the enamel. This took place at a time when it
became necessary to find a substitute for natural gas in firing a
direct-fired enamel furnace. The natural gas had never given
trouble from sulphur scum. The first producer gas used was
made from a coal which contained 2.79 per cent sulphur and the
second test was made with a gas made from a coal containing 0.77
per cent sulphur. In both cases there was a very definite sulphur

AMERICAN CERAMIC SOCIETY. 647

cum which did not appear immediately on cooling but which
ras very evident after 24 hours.

It is particularly interesting to note that in the several cases of
:umming — due to magnesium sulphate and the producer-gas
lei — the sulphur scum was carefully tested and identified as
xlium sulphate, a fact which confirms very definitely the state-
lents of Mr. Staley and of the other authors to whom he refers.

The Elyria Enameled Products Co.
Elyria, Ohio.

OBSERVATIONS ON APPARENT CAUSES OF FAILURE
OF LEAD GLASS POTS.1

Hy A. F. Gorton.

Introduction.

In order to appreciate the difficulty of building a satisfactory
pot for melting lead glass, it is necessary for one to examine a
number of pots after removal from the furnace, and to see for
himself the many ways in which failure is possible. When one
considers the unusual physical and chemical strains that a pot
must withstand, it seems remarkable that more than a single
melt is obtainable, yet 15 to 20 melts is only an average life, and
even now, three years after the disappearance of the lamented
German clay, records of 40 to 60 melts are occasionally encoun-
tered. A pot must be heated with intelligent care in the arch, or
preheating furnace, and this by the way is no mean task when
one considers the rather crude design of the arch, the inability to
distribute the heat properly, and the constant temptation to hurry
the pot into the furnace for the sake of maintaining produetion(!);
secondly, the pot must be transferred to the furnace, and this
cannot be accomplished under present-day methods without sub-
jecting it to a certain amount of strain, partly mechanical and in
part thermal. Once in the furnace, the pot is subjected to both
physical and chemical strains: (1) the throwing in of the relatively
cold (and sometimes damp) batch is a distinct shock, and must
result in the formation of tiny surface cracks in the bottom,
which widen as the glass works into them; (2) there is an erosion of
the clay due to the boiling action and wash of the molten glass;
(3) the hydrostatic pressure of the glass charge is sometimes
sufficient, during the melting period, to cause the side walls to J
bulge outward or "belly." These are physical actions. The
actual corrosion of the clay by the glass and especially the slag-
ging action of iron may be cited as chemical causes of leaks in
pots.

1 Read at the meeting of the Northern Ohio Section of the American Ce-
ramic Society, Cleveland, O., December 9, 1918.

AMERICAN CERAMIC SOCIETY. 649

While inspection of pots after they are hauled out of the fur-
ice usually affords some ground on which to base an inference

to the probable cause or causes of failure, there are many
lenomena noticed, at times in only one pot out of 20, at other
nes in a succession of pots, which baffle the combined efforts

glass-house manager and potmaker to clear up, and remain a
lzzle. The discussion that follows is based upon the writer's
nervations' and impressions gained in the examination of the
mains of some used pots which were broken up under his per-
nal direction at the Toledo plant of the Libbey Glass Co.,
id though the limited number of pots precludes any numerical
neralizations as to the causes of failure, still it will serve as a
tisfactory illustration of the types of failure commonly ob-
rved. Twenty-two pots in all were inspected, and it was found
lat the general defects could be listed as follows :

1 . Pitting and corrosion of the bottom, resulting, in a majority
j cases, in a leak at one point or another — particularly marked
; the knuckle (or intersection of the walls with the bottom).

2. Cracking of the bottom, with no sign of corrosion save in
le jaws of the crack. (It seems that the glass does not attack
le bottom unless the smooth surface is broken through and the
Drous clay of the interior exposed.)

3. Crack in the side, extending clear through, causing escape
r the glass.

4. Side walls scored or furrowed from metal-line down. This
:tion probably starts from a rather deep crack.

5. Hole through the side. It may be found on either side,
Dout twelve inches or more above the bottom. Horizontal
penings, especially near the bottom, are not herein included.

6. Sides bulged. In some instances the pot is strained at
le bulge to such an extent that a crack develops, letting the glass
>cape; at times a rip in the back accompanies the bulging. If
le clay is not sufficiently refractory, the pot will squat or col-
ipse, and this condition was not uncommon some years ago,
rhen the necessity of seeking new mixtures of raw clays and new
roportions of clay to grog was forced upon the potmaker.

7. Hole through the back. This occurs sometimes at a
eight of 12 to 15 inches, at times much lower.

650 JOURNAL OF THE

8. Crack through the back. These cracks, like those in the
side, are always \ ertical.

9. Cracking under the hood, resulting in a leak. This fault
was noticed in 4 of the 22 pots, and is not altogether to be laid
to carelessness on the part of the potmaker, as a mechanical
strain is unavoidable at this point, due to the construction of the
pot, which is not lessened by the fact that the hood is the coolest
part of the pot in the furnace.

10. Cracking and caving in of the crown. The construction
of the pot-furnace is such that the heat strikes against the crown
and back of the pot. The temperatures reached at these points
are undoubtedly as high as 27oo0F.,as the clay is often blistered
(overtired) on the surface, and beneath the surface has reached
a glassy or "vitrified" stage. There is a gradual change in color
from the back wall to the bottom, the black shading off into a
reddish buff, and this into a very pale buff or cream color in the
middle of the bottom. Corresponding to this change in hue there
is a change in porosity, for the parts described, from 5 per cent to
25 or 28 per cent. From the value of the porosity, it is quite
easy to calculate the maximum temperature attained at the
point in question, after the temperature-porosity curve has been
determined experimentally for the pot-clay. The effect of this
severe heat is to flatten the crown, but it rarely causes a collapse.
Cracks in the crown result probably from strains set up in drying
the pot. The crown, being the thinnest portion of the pot (2 to
3 inches as compared with 4V2 inches in the bottom and walls),
dries most rapidly, and, it seems, also cracks most readily. Then,
too, it is hard for the potmaker, in shaping the crown, to pack the
clay as firmly as in the case of the walls, upon which he can
exert the unimpeded strength of both hands.

From glass-house reports covering a period of several months
or a year, it is possible to gain an idea as to which types of failure
are most frequent. The writer has gone over reports from the
Libbey Glass Co. for the four months June-September of this
year, in which a total of 379 pots are listed, and finds that 81
per cent broke in the bottom, 1 1 per cent in the side, and 6 per
cent in the back. Of the 22 pots above referred to, 18 broke in
the bottom, — or a percentage . of 81 — two in the back, and two in

AMERICAN CERAMIC SOCIETY. 651

» side. The agreement of the percentage of bottom breaks is
coincidence, but it shows clearly how the vast majority of pots
1. In the following discussion an attempt will be made to
e the various possible causes of such leaks, with a view to ex-
Lining the difficulty, under present-day glass-house procedure,
remedying these defects through alteration of the ingredients
the proportions of the pot batch.

Bottom Leaks. — Taking up first the case of bottom leaks,
vas noticed that 8 pots failed through cracking, and 10 through
rosion. Of the former, two gave one melt each and one gave

melts. In the last case there were two cracks running from
nt to back, exactly over the points where the pot is supported

the fork, when it is brought from the arch to the furnace,
ice the explanation is evident. Since only one of the 8 pots
s of long life (28 melts), it seems likely that cracking in such
;es is due to strains set up (1) in drying the pot, (2) in the pre-
iting period, (3) in transferring it to the furnace. As to the
it cause, it may be stated that when the bottom cracks open
drying, the direction of the crack is always from side to side,
?., at right angles to the length. The writer, where possible,
:ed the direction of the cracks in the broken pots, and in the
.jority of cases they ran from front to back. This applies only
long cracks. Jenkinson and Marson,1 from observations made
the works of the Edinburgh and Leith Flint Glass Co., Edin-
rgh, state that insufficient preliminary burning in the arch
,y result in a sudden shrinkage after setting in the furnace,
I as the bottom is much cooler than the sides, owing to the
lductivity of the bench supporting the pot, the excessive shrink -
1 of the walls may force the bottom upward into a concave
ipe, and thus crack it. While the writer has seen no trace of
s trouble, it may be well to point out the possibility of a differ-
: source of trouble: distortion of the bottom by an uneven bed.
is also was not noticed in the present work, as the bottoms of
of the pots were plane.

Corrosion. — Turning to the next point under discussion — fail-
: of pots due to corrosion — mention was made of the fact that

pots of the 18 showed corrosion distinct from that which

1 J. Soc. Glass Tech., Sept., 1918, p. 175.

JOURNAL OF THE

accompanies a mere crack. That is, an appreciable volume q
clay had disappeared, mostly from the bottom. These pots hai
given from 2 to 67 melts, which is surely the extreme range of life
In 8 pots of the 10, close scrutiny revealed the presence of iron
which had worked into the crevices, slagging the clay and openinl
the- way for the leak. It was also noticeable in the form of ven
dark green or black glass lining the sides of the hollow or crack
Invariably the iron found the lowest point in the bottom. Thl
was true also of metallic lead, which occurred in only 4 pots. Ir
all four pots black glass was present, and the drops of lead wen
always imbedded in a dark mass of slagged clay. This seems t(
show that there is a sort of reduction of lead silicate by the iron
forming ferrous silicate and metallic lead. Evidence of corrosior
by the lead itself was lacking, since what was observed could b(
due to the action of iron. Greaves-Walker1 states that in th<
reduction of ores in the assay crucible, lead-bearing ores are ven
corrosive on crucibles high in silica. As the Gross Almerode cla}
is very silicious, this may account for the opinion commonl;
expressed by glassmakers that the German clay dissolved rapidb
and evenly in the glass (though without throwing stones).

Effect of Iron. — There are two ways in which iron may ente
the glass: from the glass batch, and from the pot itself. As I
the former, it should be pointed out that iron may enter as th|
metal itself, introduced purely by accident, it is to be hoped, o,
as a legitimate mineral constituent of the batch. (This statemenl
applies with equal force to the clay.) As to the former, it ma
frankly be admitted that under present conditions iron ma
occasionally be introduced, whether as scale or broken parts froi
grinding and mixing machinery, carelessness on the part of wort
men in leaving nails, knives, bolts, etc., around wdiere the;
can be swept among the raw materials, or in other ways; an(
once in the raw materials, it escapes detection until the pot
broken up. It has been the writer's experience that iron froi
such sources is far more likely to occur in the glass mix than i
the pot clay. In the shaping of the pot, the clay is applied i
such thin layers that detection of foreign matter is easy. Ti
prevention of iron in the glass mix can be accomplished only I
screening and magnetic separation at the last possible moment befo
1 Trans. Am. Ceram. Soc, 12, 54 (1910).

AMERICAN CERAMIC SOCIETY. 653

troduction into the pot. That such precautions are ignored by
f majority of glass-houses is an established and deplorable fact.
The percentage of iron as a constituent of the raw materials
of course very low in the case of the soda-ash, niter, sand,
i lead and other substances entering into the glass, as corn-
red with the pot clay, which may have an iron content as high
2 per cent. From a study of the burned clay it is clear that
ly a small fraction of the iron occurs as a more soluble form
jlphide and oxide), the greater part being combined with the
iv, presumably as a complex silicate, and evenly distributed
rough out the mass; for the "iron spots" are widely scattered
lereas the clay body is evenly tinged with a buff color. Analyses
the raw clay rarely show traces of sulphur. Now, if a con-
lerable portion of the iron occurred as sulphide or oxide, the
ler surface of the walls of the pot would corrode evenly and to
marked degree, and naturally the glass would always show a
sen color. It is a fact, however, that few spots dissolve markedly
the walls, and, on the other hand, in a number of cases where
:onsiderable volume of clay had been removed from the bottom,
ere was no trace of anything but clear glass. When masses of
ty are torn bodily from the surface and float in the glass as
tones," it always happens that the clay is imbedded in white
iss. If iron were to pass into the glass as the clay dissolves, it
comes difficult to explain those cases of marked corrosion where
>n is not observed. As the dark glass is always observed at the
vest points of the bottom, it is evidently dense enough to resist
ing swept upward by the boiling of successive melts. The
pothesis that iron is dissolved out of the clay is not supported
' the evidence at hand. The extremely black color of the glass
the fissures, the slagging of the clay around the leak, and the
esence of metallic lead all point to the presence of metallic
>n, which must have entered in the glass mix.
In the absence of iron, corrosion seems to be a combination of
3sion and chemical action. This begins with the formation of
perficial crevices, set up probably when the batch is thrown
As the chemical processes of melting proceed and the gases
e evolved, a boiling action develops together with a motion of
e glass in currents which sweep over the sides and bottom of

654 l« >URNAL OF THE

the pot, these currents rising along the back and ides, which ar
much hotter than the bottom. This motion tears away portion
of clay lining the crevices into which the glass has worked, givin
rise in time to pitting or honeycombing, which is especially markc
in the knuckle. The clay to a great extent passes into the glas
or "dissolves," thus doing no harm, or rises to the surface as
scum which can be skimmed ofT. This is not true in other case:
where the clay remains in the glass as "stones," and the opaqu
particles are discovered in the finished ware. »Solution of th
clay in the glass results in the formation of feldspars, and, th
glass being normally unsaturated with respect to alumina, th
solution may proceed until the bottom of the pot is less than a
inch thick, without actual leakage occurring in the worn ares
"Stones" may be explained by assuming that the pot has bee
insufficiently heated in the bottom, which is in a soft-burned c
"biscuit" state, and that erosion is very rapid, the fragment
passing into the glass being so large that they cannot dissolve i
the fining period, or that the glass dissolves some of the clay an
then becomes saturated with respect to alumina. All availabl
evidence shows conclusively that soft-burned, porous botton
corrode excessively, while pots which have been burned to a dens
structure offer great resistance to solution. This has been show
in the laboratory by melting a rather corrosive lead glass in 5-inc
crucibles, a single melt sufficing to show marked or complel
penetration of the bottom of the soft-burned crucibles, while tr
ones heated to a higher temperature showed a straight, whil
line of separation between the bottom and the glass. In tr
present work it was noted that all pots giving 5 melts or less wet
spongy and a light buff or cream color in the bottom. In fact,
is not possible under present conditions to burn the bottom i
the arch to anything like the temperature needed to produce
dense, resistant structure. Travers1 has shown that this difficult
may be overcome by allowing the pot to remain in the furna<
at 13500 C for three days before filling, a solid wall being built t
9 inches in front of the pot, so as to ensure even heating. Th
practice would do much to remedy conditions in this countr
though American glassmakers would undoubtedly object to tl
1 /. Soc. Glass Tech., Sept., 1918, p. 171.

AMERICAN CERAMIC SOCIETY. 655

:s in production such a course would entail. With the idea of
ating the bottom to a higher temperature, Jenkinson and Mar-
11 have tried supporting the pot in the furnace on lire brick,
t found that the bottom warped. At various times, glass
ikers have tried burning the inner surface of the pot by intro-
cing a gas- or oil -burner, but this practice was discontinued
cause it was impossible to spread the heat over a large area.
Sillimanite. — The formation of sillimanite (Al20;j.Si02) has been
vanced by many investigators in explaining the resistance
ered to corrosion by clay which has reached the "vitreous"
tge. An exhaustive investigation of the formation of sillimanite
d related minerals in glass pots and glass tanks has been made
Wilson,2 who finds that, normally, sillimanite is developed only
the inner and outer vitrified surfaces of the pot, and not to any
tent in the bottom. He also states that, with an aluminous
y, the boundary between the glass and the vitrified clay is
ry sharp, the hard clay surface being full of sillimanite needles,
lereas a silicious clay is attacked in an uneven manner, the
undary being irregular and pitted. The life of the pot then
pends on the rate of solution of the sillimanite layer; in the
mer case the glass adjacent to the pot is oversaturated with
;pect to alumina, whereas in the latter case it is unsaturated,
le argument applies with equal force to tank blocks, particu-
ly flux-line blocks. Since the heat inside a tank furnace is
ver sufficient to vitrify the blocks, it is imperative that all tank
)cks be burned to a dense structure in the kiln. Travers3
ints out very clearly that when the clay is vitrified merelv by
itact with the molten glass, the layer of sillimanite formed is
ceedingly thin, whereas that formed in a thoroughly vitrified
t is rather thick. Corrosion in the former case is relatively
pid, especially in the bottom, which is not vitrified by the exter-
1 heat. Consequently it is useless to vitrify a pot in the manner
scribed above; moreover, it is doubtful whether much is to be
ined by lining the pot with aluminous or similar material, be-
use of the danger of such a lining cracking or wearing away in
Dts, exposing the porous inner clay.

1 Jenkinson and Marson, Lor. cit.

- G. V. Wilson, J. Soc. Glass Tech., Sept., 1918, p. 207.

3 Loc. tit.

656 JOURNAL OF THE

Holes. Concerning the remaining points enumerated above
only a few comments need be made. The- question of hold
developing in pots is an interesting but puzzling matter. It haj
been suggested that corrosion begins around a particle of iron
(pyrite?), and the glass, having penetrated the surface, easily
attacks the less resistant clay beneath. The trouble is one tli.it
at times crops up only in isolated cases, at times in a succession
of pots, and is on the whole rather a mysterious matter. Of the
22 pots inspected, one had a hole through the side and one a holt
through the back. The glass that escaped was clear, and Hurt
was no discoloration of the clay around the leak. On the other
hand, of 4 pots developing cracks in the side or back, all showed
traces of black glass. The theory that holes are due to metallic
iron imbedded in the clay would be more impressive if discolora-
tion and slagging of the clay beneath the hole could be detected
in every case.

Deformation. — The question of softening and bulging of the
pot in the furnace is always important, though, as previously
stated, this defect is not usually serious, but occurs when the
potmaker is forced, through inability to obtain a raw clay or
deterioration of his regular bond clay, to experiment with new
material and alter the proportions of raw clay to grog. Certain
practices and conditions in glass works are productive of bulging
(1) failure to take down the stopper at the proper moment in the
fining period, (2) "forcing" a pot in the hope of remedying stones
color or other defect in the glass, (3) excessive heating of a few
pots due to uneven distribution of the heat within the furnace
As a means of judging the resistance to softening, the writer ii
both routine and experimental work has employed the warpage
test, and has found it to be a direct and satisfactory method. Th<
maximum temperature employed in this test was 27000 F., rathei
than the lower figure of 25000 F., which may be taken as the
upper limit for the temperature of the glass. In the light of 1
previous discussion, it is probable that certain portions of th<
pot reach a temperature even higher than 27000 F. This sur
mise is based on the observed color and porosity of pieces takei
from the crown and back wall of a number of pots. The porosit;
is found to increase in the direction of the bottom, which is alway

AMERICAN CERAMIC SOCIETY. 657

ry porous as compared with the "vitrified" portions further up.
lis condition is very bad, for it aggravates the natural tendency

the glass to corrode the bottom. To correct it, either of two
urses may be followed: (1) to burn the bottom to an elevated
mperature in the arch or furnace, or (2) to make the bottom

a mixture that becomes dense at a lower temperature.
The former alternative has received some consideration. In
^ard to the second course, from the writer's experiments and
tservations of pots recently in use, it appears that there is little

be gained by adding porcelain or similar feldspar mixtures to
e pot clay, so long as the pre-heating is not sufficient to bring
•out a dense structure in the bottom. When pieces of such pots
e examined, great fissures are seen extending everywhere
rough the mass, as though the dense, whitish grains of porcelain
id shrunk away from the more refractory matrix. This condi-
i>n is due to the fact that the different materials were not inti-
ately blended, and is unavoidable when the porcelain is intro-
lced as grog. However, the writer has found that this segrega-
on of the porcelain in the form of white grains occurs even
hen the raw kaolin, feldspar and flint are individually ground
iry fine and thoroughly blended with the fire clay. Where such

mixture is in contact with the glass, vitrification extends no
irther than a depth of about a millimeter, and is not sufficient
rotection against corrosion because of the rapid solution of the
[trifled layer and the danger of a crack exposing the porous
laterial of the interior. Bleininger1 has recently shown that the
Drcelain optical glass pot may corrode as much or more than an
■dinary pot unless it has been heated to at least 14000 C before
[troduction of the glass batch.

If the bottom alone were made of a different, denser-burning
tixture, trouble would arise from the difference in shrinkage
Jtween the bottom and walls, and the possibility of cracking at
le knuckle. The fact that bottom-clay is worked stiff er than
de-clay would tend to obviate the former, but increase the latter
fficulty- Under present working conditions, the walls must not
5 made less refractory, rather the contrary; moreover, if the pot
jere as a whole more refractory than it is now, special glasses
' J. Am. Ceram. Soc, 1918, p. 21.

658 JOURNAL OF THE

requiring a more elevated working temperature could be melted.
All things considered, it is to be hoped that experiments on the

use of more refractory substances than lire clays as basi< mil'
rials in pot construction may be carried out by the intelligent
cooperation of pot manufacturers and glassmakers, and that such
attempts will lead to gratifying results.

Summary.

1. Examination of the remains of pots used in melting lead
glass shows that cracking and corrosion are the chief causes of
failure.

2. Cracks are due to strains set up by (1) too rapid heating in
the pot arch, (2) insufficient pre-heating, (3) by throwing in the
glass batch. The second cause is perhaps the most serious source
of trouble.

3. Corrosion is due in some cases to the slagging action of iron
on the clay, in most cases to solution of the clay in the fusing
alkalies and molten glass. Solution of the clay tends to color
the glass, owing to the iron content of the clay, but the tint can
be neutralized. Black glass and metallic lead found in pot bot-
toms are considered to be due to metallic iron in the glass mix.

4. Inability to burn the bottom of the pot in the pot arch ac-
counts for the rapid corrosion of the clay. The trouble might
be remedied by allowing the pot to remain empty in the furnace
for three days, but glassmakers would object to the loss of pro-
duction entailed.

5. The vitreous layer formed by heating the inner surface of
the pot with an oil-burner, or by the molten glass itself, is very
thin, and is not sufficient protection against the attack of the glass.
Lining the pot with aluminous or porcelain-like material will not
help if there is danger of the lining cracking away (in drying 01
burning) and exposing the more porous pot clay beneath. Th(
safer method is to burn the pot thoroughly at a temperature 0:
25000 F., the heat being applied directly to the bottom.

6. It appears that little is to be gained by adding porcelain a'
grog to the pot batch, unless the mixture is rendered dense befon
use. Moreover, this procedure is objectionable because it wouk
increase the tendency of the walls to bulge.

AMERICAN CERAMIC SOCIETY. 659

7. In view of the difficulties attendant upon the use of one
ay for the bottom and another for the walls, it seems desirable
1 search for some material superior to fire clay in refractoriness
id in resistance to corrosion.

In conclusion, the writer desires to thank Messrs. Jones, Hess,
arry and Nipper of the Libbey Glass Co. for their kind aid and
any valuable suggestions.

Ceramic Laboratory,
The Buckeye Clay Pot Co.,

Toledo, Ohio

AN UNUSUAL CAUSE OF SPALLING OF SEWER PIPE.

By Cui.i.kn W. Parmi:i.i:i:.

Not long ago the writer was requested to investigate the cause
of spalling of sewer pipe which occurred during the burning of
the ware. The pipe as it came from the kiln was marred by the
flaking of the outside surfaces. Usually these flakes were de-
tached. Sometimes they were almost but not quite separated
from the pipe. An examination of the ware showed that in
practically every instance a small fragment of what had been a
more or less spherical hollow body was imbedded in the detached
spall and the other portion of the spheroid was attached to the
surface of the fracture of the pipe. These spheroids were about
2 to 3 mm. in diameter. Under a magnifying glass they showed
the characteristic appearance of small geodes, that is, small hollow
bodies with the interior face of the cavity lined with small crystals.
Our opinion regarding the nature of these spheroids was con-
firmed by members of the faculty of the Department of Geology.

The occurrence of geodes in clay is probably rather unusual
but they are known to occur in certain clay deposits in Illinois,
notably in the vicinity of Hamilton. These geodes, we are in-
formed by one familiar with that locality, vary in size from a
marble to those which are six inches or more in diameter. The
larger are more commonly found and do not give rise to trouble
except for the expense of repairs to the machinery — due to break-
age if these hard nodules are not first removed from the clay.

The spalling, which we observed, accompanied the use of a
clay which is of a different character and geological occurrence
from that at Hamilton. At the latter place, the geodes are of
common occurrence while at the other locality their presence
was not known previous to our investigation.

After having determined the probable cause of the trouble, we
undertook to locate the stratum containing these nodules. A
search of the floor of the clay pit was rewarded by the discovery
of a geode imbedded in a large block of a sandstone conglomerate,

AMERICAN CERAMIC SOCIETY. 66 1

ich originally had formed part of the rock mass underlying the
y deposit. A painstakingly thorough examination of the rock
ge resulted in the discovery of a number of small geodes vary-
; in size from a hickory nut to a large bean. Together with
:se geodes was found a remarkable accumulation of more or
5 perfect quartz crystals, double ended, each not much larger
in a kernel of rice. These geodes apparently were located in a
y small area on top of the sandstone underlying the clay
iosit.

Phe spalling occurs, apparently, during an early stage of the
rning. This is indicated by the small cracks on the reverse
e of each spall and radiating from a point almost exactly above
: geode fragment — showing that a strain was developed while
: clay was at a temperature too low to have given it a permanent

Phe small geodes which we found were all lined with quartz
stals. We are uncertain as to whether the spalling was due
the popping of the geodes — due to the expansion of the air-
id cavities — or whether it was due to the expansion of the
irtz particles. Probably the former was the cause.

Department of Ceramic Engineering,
University of Illinois.

ACTIVITIES OF THE SOCIETY.
January 4, 1918.
Annual Meeting.

Plans arc being rapidly formulated for the Annual Meeting
which is to be held in Pittsburgh on February 3, 4, and 5. Thi
will be the twentieth anniversary of the formal organizatioi
of the Society, and preparations are being made to commemora
the occasion in fitting manner. Among other things 1 xinj
planned are a banquet, the elevation of a large group of associati
members to active membership, the election to life membershij
of some of the men active in the organization and early develop
ment of the Society, and the conferring of Honorary Membershi}
on one or more men whose services in the advancement of ceramii
technology has been especially noteworthy. Arrangements ar
being made to have representatives of allied scientific and tech
nical societies take part in the anniversary program.

This meeting will be of special interest also from the stand
point of the technical and practical worth of the papers and dis
cussions. The forming of Professional Divisions is arousin
great interest among the manufacturers and plant executives c
the various industries represented on our membership roll:
Papers of interest and practical value to each of these grouf
are being prepared and arrangements have been made to havi
these read and discussed in separate sessions of the Profession; I
Divisions. This will give opportunity for free discussion am
interchange of views by the men actively engaged in the particul.'jj
industry represented by the Division. More papers of the kir
indicated are earnestly solicited.

A joint session with the United States Potters' Associatio
many of whose members are aso members of our Society, is beii'.
arranged. During the same week, annual meetings will be he
in Pittsburgh by the Nationa Brick Manufacturers' Associatic
the National Paving Brick Manufacturers' Association, t
National Building Brick Bureau and the National Clay Machine

AMERICAN CERAMIC SOCIETY. 663

anufacturers' Association. Opportunity will be given for in-
ection of the laboratories of the Bureau of vStandards and
ellon Institute as well as the unparalleled variety of ceramic
mts in the Pittsburgh district.

At the business session several matters of vital importance
the management and development of the Society will come up
■ decision.

Make your plans now to attend the meeting. It will be the
•gest, most important, and most enjoyable meeting ever held
the Society. You cannot afford to miss it.

Professional Divisions.
In accord with the motion of Mr. Montgomery reported in Num-
r 8 of the Journal and recently adopted by the Board of
ustees, the President has appointed the following members
act as organization leaders of Professional Divisions.

White-ware and Porcelain C. L. Sebring

Refractories A. V. Bleininger

Glass Technology C. H. Kerr

Abrasives R. C. Purdy

Enameled Metals R. D. Landrum

Terra Cotta and Faience F. B. Ortman

Brick and Tile M. P. Blair

Members are urgently requested not to wait for an appeal from
e leaders, but to become actively interested themselves and to
Duse the interest of every ceramic man they know in the organiza-
>n of one of these Divisions.

Nominations for Active Membership.
The Board of Trustees desires that members send to any
jmber of the Board at once nominations for active member-
ip. Many of our associate members deserve this elevation,
d it is desirable that the voting power in the Society be more
dely distributed. It is to be noted that the presentation of
pers is no longer the sole basis for elevation to active member-
ip. Following are extracts from the revised rules of the Society
vering this subject:

'Active members must be persons competent to fill responsible positions
ceramics. Only associate members shall be eligible to election as active

664 I" >URNAL OF THE

members and such elections shall occui only in recognition of attainment!
or of service in the science, technology, or art, <>f the silicate industries."

"To l)t- promoted to active membership associate members must be noun
oated in writing by an active member, seconded by not less than two other

active members, and approved by the Hoard of Trustees Their nomination

must be accompanied by a statement of t heir qualifications and to be elected
they must receive an affirmative vote of the majority of the members of

the Hoard of Trustees."

Honorary Secretaryship.

DECEMBER 22, 191X
Mr. Homer K. Staley,

Bureau of Standards,
Washington, D. C.
My Dear Staley:

I am in receipt of your message to the effect that the American Ceramic
Society desires to create a new office, that of Honorary Secretary, and to
elect me to it. I am, of course, not insensible to the honor which the Society
desires to confer upon me, and greatly value this added proof of their good
will. Nevertheless, I must refuse it.

The custom of having Honorary officers is common and well established
in English and European Societies, but here it is rather anomalous. In
England, the Honorary officers have little part in the actual achievements
of the organization, the work being done by the active officers, but the position
tills a useful purpose nevertheless, as there are always members of the upper
class whom it is desirable to recognize.

I should be loath to see this custom, which belongs to a social structure
quite different from ours, become popular in the United States, and still
more unwilling to see myself figuring in such a capacity. When the exigencies
of the war are ended, if I find that I can take up the threads of my former
life again, the only status I should care to have in the Society would be that
of an active worker, where I could mingle on terms of exact equality with
everybody, and where I would be unhampered in making any views and
purposes felt.

I feel sure that you can make the Society see, after they have given the
matter a little more deliberate consideration, that the course which I propose
is really the safe and democratic one and is for their best interests.
With kindest personal regards,

(Signed) Edward Orton, Jr.,

Lieut. -Colonel, M. T. C,
United States Army.

JOURNAL

OF THE

MERICAN CERAMIC SOCIETY

A monthly journal devoted to the arts and sciences related to
silicate industries.

. 1 October, 1918 No. 10

EDITORIALS.

IMPORTED CLAYS.

lith the signing of the armistice there appeared to be a eessa-
of the agitation of many months past in reference to the
ortation of the clays of different kinds. Undoubtedly with
closing of hostilities many of our manufacturers began to
: with longing eyes to a fresh and adequate supply of the im-
;ed materials, especially those used in the manufacture of the
te burning pottery products. Nor were they doomed to
ppointment as we have already learned of an increased flow
he imported clays.

hat the efforts of our geologists and laboratories have not
l fully rewarded in the location and development of new and
:mate deposits of white burning clays of first quality is to be
etted. Let us hope that there will be no let-up in their
rts and that the field and laboratory searches will be con-
;ed.

•oubtless the situation as regards the importation of clays for
able, pencil and glass pot manufacture will remain the same
some time to come. American clays in glass pot mixtures
ear to be giving very good satisfaction and the production of
s for this purpose appears to be on a well established basis,
results secured with American clays in graphite crucibles
showing steady improvement, and it is to be hoped that our
ible manufacturers will never again be dependent upon the
ign deposits for this highly important commodity in their"
istry.

666 AMERICAN CERAMIC SOCIETY.

THE POTTERY INDUSTRY.

When viewed as a whole, there has not been a marked chaugi
in the quality and general nature of the pottery wares manufacture

in this country during the years of the war. It is true that m
have long excelled in the quality of certain of our pottery man!
faetured products, especially of a structural nature such a
electrical porcelain, sanitary ware, etc. In a general way, how
ever, the hulk of our wares is of the same quality as before th'
war.

Our pottery manufacturers still cling to the use of ball clays-
True, the elimination of ball clays from body mixtures involve
more or less radical changes in methods of molding, buniin
temperatures, etc. So long as the ball clays are used, howevei
it will not be possible to produce wares which compare favorabl
in color and cleanliness with the hard porcelains of Continents
Europe. We do produce in limited quantities wares which ar
not inferior in appearance and quality to the best of the inij
ported wares. Our wares of this kind, however, must be class©;
as art wares and cannot compete in price with the bulk of th
table wares which were imported in large quantities previous t
the war and sold at amazingly low prices in competition with th
American-made pottery.

Will our manufacturers be satisfied to return to the forme
conditions of competition or will their efforts to raise the standar
of quality be stimulated? It is to be hoped that the techniqu
of this branch of the pottery industry will keep pace with th
advancement made in some of the smaller and less importat
branches.

ORIGINAL PAPERS AND DISCUSSIONS.

A DISCUSSION OF TERRA COTTA DRYERS.

I3y Ellis Lovejoy, E.M., Columbus, ( >.

As we recall the various terra cotta dryers we have examined
d designed, the thought comes to us that the manufacturers of
ra cotta have not come to any system in dryers ; that the dryers
>d differ widely in character and efficiency; that the ware, built
from several clays and grog, will stand severe treatment in
? simple shapes and only the large and complicated shapes
mire more careful treatment; and finally that increasing com-
tition will require the manufacturers to adopt more economical
thods of drying and burning.

fhere are in use several ordinary waste-heat car tunnel progres-
e dryers. The waste heat progressive dryer holds a prominent
ice in the clay industries. It uses the heat of cooling kilns and
it of engine exhaust, and, in consequence, it is economical in
I. The dryer conditions seem to be ideal. The ware enters
lumid low temperature atmosphere and as it advances, the
tnidity decreases and the temperature increases, until finally
ir the exit it is subjected to upward blasts of hot air through
: floor ports. It is desirable in the manufacture of terra cotta
it each lot of ware come through about the same time and where
;re are a number of tunnels this condition is fairly well met.
e operation of the dryer is not so ideal as it would appear,
long as the operation is continuous the conditions in the tun-
s are constant, but continuous operation is not customary.
ere is a forward movement of the heat and a drop in humidity
the tunnels during the night when the ware is not moving,
I a corresponding retreat in the daytime when the tunnels are

668 JOURNAL OF THE

being emptied and filled. This variation in the tunnel rondi-
tions may be withstood by the margin of safety in the drying
behavior of the ware. The tunnel conditions vary much more
widely over holidays and some wares may not stand this varia
tion, but stopping the fans during all or part of the period may
give sufficient control. Thus far, however, the progressive tunnel
may be said to meet the requirements.

There is a more serious objection. It is impractical to so load
the ears with terra cotta as to closely fit the tunnels. Almost in-
variably there is excessive space between the tunnel roof and the
top of the ware and the air rises into this space. vSince the air
movement beyond the hot air zone in the tunnel is horizontal, the
top of the ware, or the top ware, is subjected to maximum drying
conditions while very little drying progress occurs in the bottom.
When the cars reach the hot zone the bottom green ware is sub-
jected to the severe conditions imposed by the blasts of hot air
coming through the air ports and it is to this ware in this zone
that great damage is likely to occur. It is readily seen that the
top ware is safe in the hot zone because of its advanced condition
and its position relative to the air blasts. It frequently happens
that the top ware is damaged in the early stages near the receiving
end and this is often attributed to a lack of humidity, but it is
our opinion that it is more often due to unequal drying.
The users and advocates of the progressive tunnel may say that
the ware readily withstands these conditions and that the dryer
is satisfactory. We wonder how far this is true. If there is no
loss in either dryer or kiln there can be no question in regard to
the efficiency of the dryer and the operators are to be congratu-
lated on the strength of the body which they have developed, but i
this is not conclusive, and one is not warranted in adopting such a I
dryer as the best for the product in general. In other lines of
ware we have too frequently seen the weakness of the progressive >>
dryer to be convinced that it is the best type for the manufacture
of terra cotta.

In drying other wares we have found a decided improvement
both in the rate of drying and in the per cent of perfect output by I
extending the hot air duct nearly or quite the full length of the
tunnel and thus securing up-draft through all the ware. At first

AMERICAN CERAMIC SOCIETY. 669

nee this seems more severe than the original type, and on the
ole it may be so, but the ware is subjected to more uniform
ditions. The temperature of the air is lowered by wall and
und absorption as it advances in the lower duct, and with
duated floor slots the conditions at the car receiving end may
materially less severe than at the delivery end. We still have
forward movement of air in the tunnels with increasing humid-
and there will be decidedly more circulation in consequence
the upward currents into the horizontal current. We have
1 terra cotta regularly subjected to decidedly more severe
itment, and while this ware may have been exceptional in its
ing behavior, yet the fact that it is a prepared body heavily
f*ged leads us to believe that it will generally stand more severe
ing tests than many of the common wares made from a single

'he chief weakness of the usual progressive tunnel is lack of
ulation. We dry and harden the top ware without getting
ing conditions down into the mass of ware. If the ware will
id the direct up-draft heat the full length of the tunnel as above
jested, an economical dryer for terra cotta is a simple prob-
, but if not then we must improve the operation of our present
e by better circulation, or in other words, break up the hori-
tal current of air in the top of the tunnel and drive it down
mg the ware.

'he value of circulation is remarkably demonstrated in recent
srs nor is the idea new.

I seems to us that in the progressive dryer, the operation
Id be in two stages. Make the delivery (hot) end as in the
il progressive tunnel, then midway, or less than midway, of
tunnels put a cross duct on top of the tunnels, similar to the
er-ground hot-air duct at the delivery end, with an opening
>ugh the bottom of this duct (tunnel roof) into each tunnel,
w the air from the tunnels by means of a fan and return it
he tunnels through a second cross duct and lateral ducts on
of the several tunnels. In other words, duplicate on top of
tunnels from the center to the receiving end, the duct system
:he delivery end, except that the air for the receiving end
Id be drawn from the tunnels themselves instead of from

670 J< >URNAL OF THE

cooling kilns or through steam coils. The partially to nearly
dry ware in the delivery end will be subjected to upward blasts of
hot dry air. Prom the center to the receiving end the green ware
will be Subjected t«> downward blasts of lower temperature highet
humidity air in which the temperature decreases and the humiditf
increases toward the end. In such a construction we retain all
the ideal features of the progressive tunnel and overcome the very
serious horizontal over-draft difficulty without any increase in
fan equipment of the modern two fan dryer.

A progressive tunnel equipped with steam coils only and mak-
ing no use of waste kiln heat is, or soon will be, economically out
of the question. In our opinion the most essential economy in
the future of the business is the use of the heat from the cooling
kilns for drying even though the future factory makes use of tht
continuous or car tunnel kiln.

The box systems in use in several factories, it seems to us, an
far from satisfactory and are not economical. The ware is sei
in a closed room. Blasts of air are turned into the room and
by one means or another, directed to an exit. The treatment i:
severe, and the heat waste excessive. The air volume and con
sequent heat supply is practically constant, while the moistun
removed passes from a maximum to virtually zero. At the en<
of the drying we are heating a large volume of air and forcing i
through the dryer without any return whatever. A better opera
tion is to return the air to the fan, replacing some loss with fresl
air. Thus we get the humidity conditions and at the same tiro <
conserve the heat. The usual equipment is the steam coil witil
exhaust steam in the day and live steam at night. So far as w
know, waste heat from cooling kilns is not used, perhaps becaus
of the greater ease in regulation of the temperature and als
because the profits of the business have not insisted on economie;
The introduction of kiln heat into a factory building not fire
proof involves a questionable fire risk although a number of sewe
pipe factories have made such use of it.

The humidity treatment is nicely carried out in the regenerc
tive type of tunnel dryer. The tunnel is two storied. In the uppe
tunnel, or duct, are steam coils, with a fan at one end. The low<
tunnel is filled with cars of ware. The two tunnels are connecte

AMERICAN CERAMIC SOCIETY. 671

each end. When the cars are placed, the doors closed and t lit*
1 started, the fan picks up the air from the lower tunnel, drives
through the coils to the opposite end of the upper tunnel, then
jvn into the lower tunnel and back among the ware in the lower
inel to the fan. Thus we have a complete cycle repeating ad libi-
n. It is necessary, of course, to provide a vent for the escape of
ne of the saturated air and replace it with fresh air through an inlet
the fan, and this may be extended to the extent that all the
ake air is from the outside without any regeneration. T lu-
mber of such dryers is too limited to draw any conclusion. It
in a class with the regenerator box dryers with the handicap
multiple fans and horizontal air movement. Steam supplies
• heat, and live steam so far as we know, but there is no reason
y exhaust steam may not be used. It is not practical to use
ste kiln heat in connection with regeneration and natural humidity
*ept by the use of an economizer equipment which would be ex-
lsive in initial installation and maintenance. There is, however,
excess of waste kiln heat and no occasion for regeneration except
get natural humidity. Humidity could be obtained in other
ys than from the ware, but it requires regulation and risk not
Solved in the natural operation. It is inconceivable, especially
the tunnels are short, that moisture taken from the ware in
:hange for heat could saturate the air to the extent of the dew
nt, whereas careless artificial humidation might easily saturate
excess of the dew point and result in considerable damage to
: ware.

We have in this regenerative tunnel dryer the same difficulty
horizontal air movement over the top of the ware without
isfactory circulation around the ware as in the progressive car
inel and in greater degree in that there is no upward movement
in the delivery end of the progressive tunnel. The horizontal
►vement can be broken up by a duct between the upper heating
inel and the lower drying tunnel, with ports into the latter.
ius the air from the upper heating tunnel would return through
s delivery duct and be blown down, on the ware, through the
iduated ports in the under side of the return duct. As stated
me, we do not believe that the air naturally could be saturated
excess in the original construction, but should it be possible,

67:

JOURNAL OF THE

tlu use of the return duct delivering hot air at intervals from end
to end of the tunnel would greatly lessen tlu- possibility.

The return duct could be eliminated and graduated ports In-
placed in the division floor between the two tunnels and thug
simplify the construction, provided the steam coils are al one end
of the upper tunnel.

Fig. i.

Another idea would be to place the return duct under the lower
tunnel floor and blow the air upward among the ware, thence
horizontally to the end of the tunnel and into the fan, but this
involves an awkward connection between the heat tunnel and
delivery duct.

The only advantage that we can see in the double tunnel in
its original form over the usual progressive tunnel is that it may
be limited in length and thus better adapt itself to many factory
layouts, and in that the humidity conditions are probably more-
satisfactory. Its fuel cost in live steam and multiplicity of fans
give a balance in favor of the progressive waste heat tunnel.

AMERICAN CERAMIC SOCIETY. 673

A recent innovation in tunnel dryers, not as yet, however, applied
> terra cotta, has a wide tunnel with three under-floor ducts as
lownin the sketch. (Fig. 1.) The two outside ducts supply the
Dt air from the cooling kiln and the central duct is the exhaust,
he hot air rises among the ware on pallets over the side ducts
id passes down among the ware over the central duct. The
:ntral duct could be used for hot air and the side ducts
r exhaust or two ducts could be used instead of three, —
ie for hot air and one for exhaust. If all the ware will stand
le direct heat, that slowest in drying could be placed over the
Dt air ducts and the quick drying ware over the exhaust duct.
nee the outgoing air would have greater humidity than the
coming air, the ware over the exhaust would be retarded in its
-ying. Another view would be to place the safe drying ware
/er the hot air ducts and the difficultly drying ware over the
chaust duct. This would require a longer drying period but
sures greater safety, and in some degree gives humidity for the
ore difficult ware. If a more complete humidity operation were
quired, the moisture would have to be introduced into the in-
>ming hot air and reduced as the drying progressed.

One's first thought would be the possibility of short circuiting
om the hot air ducts to the exhaust duct, but this would be easily
/ercome with the foot-pallets in use in a number of factories,
he tunnels would be roomy and satisfactory for motor trucks,
he danger of short circuits with cars would be greatly increased,
id it probably would be necessary to have a curtain wall between
ie cars which would insure the full up- and down-movement,
he feature of the dryer is the circulation, and there are no
strictions in regard to length, except that excessive length
^creases the chances of getting uniform exhaust from end to
id. This problem is less difficult and more easily solved in a
yer than in rectangular kilns where we have in some installa-
ons end-draft up to ninety feet in length.

Floor drying as in sewer pipe factories, with covers to get hu-
idity, increases the floor space required, heats the working space
iduly and increases the humidity therein, and finally requires
ve steam in a large measure. It has the advantage that the
are can be watched and the drying of individual pieces controlled,

674 AMERICAN CERAMIC SOCIETY.

l)ut the manufacture <>f terra cotta is no longer the modeling of
lions rampant and recumbent; of decorative friezes and capitals.

The l)nlk of the ware is in small pieces duplicated indefinitely and

the need of an independent dryer to take the ware from the floor
is imperative.

Ducts from the kilns or boilers under the pressing floors are in
use but in our opinion are questionable, and blowing hot air into

the press room at night from kilns or steam coils is not a sails
factory drying system. "Better working conditions" are the
basis of many labor disputes and often times justly so, and factory
operations must meet these conditions. Where it is necessary to
hold the ware on the floor until it is fully set (leather hard) it
would be better to provide a steam heated slatted floor for this
preliminary drying, or in case it is not desirable to move the ware
at all, then after filling one section of the floor, move the press-
men to another section and turn steam into the pipes under the
first section, or use steam only at night. It is not our purpose
to discuss sanitary factory conditions which have no bearing on
subsequent drying, except as the factory is the dryer. We are
of the opinion that a properly developed independent dryer with
suitable control will eliminate much of the hot floor press room
drying.

The wide divergence in dryer types indicates that terra cotta
manufacturers have not solved the dryer problem.

The essentials are waste-heat, thorough circulation and con-
trollable humidity. We do not know of any dryer which includes
all of these. In some instances humidity is of less importance.
This leaves waste-heat and circulation to be combined, yet sim-
ple as the problem is, we have not seen the solution of it in opera-
tion in any terra cotta factory.

HE COMPOSITION OF CHINESE CELADON POTTERY.

By J. S, Laird.

The raw materials used in the modern ceramic industry in
hina, together with the compositions of the bodies, glazes, and
)lors employed, have been studied very thoroughly by the cer-
mists of the Sevres potteries, Ebelman and Salvetat,1 in 1850,
id Vogt2 in 1890. The latter investigator has confirmed his
isults by preparing from European materials bodies exactly cor-
;sponding to the Chinese, and has reproduced many of the famous
azes, including crackle ware and flamed red or sang-de-boeuf.

Our knowledge of old Chinese pottery is much less complete,
ery few specimens have been examined scientifically as their
alue as works of art and antiques has discouraged the destruc-
on incidental to chemical analysis. In the hope of obtaining
aluable data on this subject an examination of the sites of the
icient kilns at King-to-chen was undertaken a few years ago by
[r. Charles L. Freer of Detroit, the well known collector and
athority on Chinese pottery. Internal troubles in China have
p to this time made excavation of the kiln sites impossible. Two
teresting specimens were, however, obtained and examined anal-
tically. These partial results are reported in this article.

Specimen 1. — This specimen consisted of a lump of glaze which
ad dripped off the ware in firing and solidified on the kiln floor,
fhile it was undoubtedly old, its exact date has not been fixed.
: was a solid, homogeneous lump, in color bluish green or "cela-
3n" with reddish streaks. Where it had hardened on the kiln
x>r it had picked up particles of sand. These particles were
irefully ground off the sample taken for analysis.

Specimen 2. — The second specimen obtained was a small plate
■ saucer about five inches in diameter. It probably belonged
» the Yuan period and showed its antiquity by a marked weather-
g. The body of this plate was fairly dense and fine grained,

1 Ebelman and Salvetat, Ann. chim. phys., 313, 257.
3 Vogt, Bull. soc. encour., 00, 5.^0.

,,;<> JOURNAL OF THE

but soft enough to be easily scratched by a knife. It was light

colored, inclined to buff with reddish streaks. Its composition
and appearance would class it as a fire clay and not a kaolinitic
body.

All but the foot of this plate was covered with a thick opalescent
bluish green glaze showing a small uniform crazing. It was
sufficiently opaque to completely hide the body of the ware. A
blotch, shading in color through bright red and green to black,
marked a spot where apparently a cinder had fallen on the ware
during the firing. Sufficient of the glaze for analysis was removed
by careful chipping with a small chisel, and thanks to the thick-
ness of the glaze it was possible to obtain a sample entirely un-
contaminated by fragments of the body. It was not possible to
obtain separate samples of the red or green portions.

The analyses of these two glazes and of the body of the plate
are given in Table i.

Table r.

Glaze 1 (lump). Glaze 2 (plate). Hody (plate).

Si02 72

AI2O3 10

Fe203 o

FeO 2

CaO 8

MgO 1

K,0 3

Na20 1

52 per cent 66.40 per cent 63 . 20 per cent

59 13-88 28.30

41 ° 30 3 47

02 1 . 19

48 1 1 58 2.05

24 1 .61 trace

77 4-25 198

48 1.84 0.37

MnO trace trace trace

Ti02 trace trace trace

P206 0.21 trace

100.72 101.05 99 37

Except for the presence of iron, these glazes are closely related
to the cone 4 type of whiteware glazes and are quite calcareous,
No. 2 being even higher in CaO -f- MgO than the cone 4
formula.

Color of the Glazes. — It will be noted that the only coloring
oxides present in the glazes in appreciable amounts are oxides of
iron, particularly ferrous oxide. The presence of the oxides com-
monly employed to give blue or green colors — copper, cobalt and
chromium — could not be detected bv the most careful tests. When

AMERICAN CERAMIC SOCIETY. 677

irticles of these glazes were remelted in an oxidizing atmosphere
ey became almost colorless. Vogt1 found that the blue-green
lor of modern celadon glazes was developed only when the ware
is fired in a strongly reducing atmosphere, otherwise a pale
How was produced. Compounds of ferrous iron must therefore
ive given the color of these glazes, the peculiar opalescence being
le probably to the high percentages of lime and alumina.
Comparison with Other Reports of Celadon Glazes. — Meyer2
s reported analyses of the body and glaze of an old celadon
,se brought from China, and Ebelman and Salvetat3 and Vogt4
alyzed modern celadon glazes and the bodies on which they

Table 2.
Chinese bodies used with colored (celadon) glazes.

1. 2. 3. 4.

Ware. Yuan. Old Chinese. 1850. 1890.

Analyst. Laird. Meyer. Salvetat-Ebelman. Vogt.

5i02 63.20 69.51 69.O 69.50

\l2O3 28.30 22.72 23.6 23.IO

?e203 347 184 1.2 2 . 00

raO 2.05 0.27 0.3 0.16

tfgO trace 0.36 0.2 0.22

^O 1.98 4-79 3-3 3 -72

^0 0.37 0.87 2.9 1. 12

99.37 100.36 100.5 99 82
Celadon glazes used on these bodies.

)iC>2 66.40 64.98 72.00 70.2

U2O3 13 .88 14 33 6.00 14. &

fe203 0.30 1.39 2.50^ 1.5

feO 1 . 1 9 .... ....

"aO 11.58 10.09 10.40 8.1

/IgO 1. 61 1 55 trace 0.2

:s° 4 25 561 / 3.2

Ja20 1.84 0.81 \ y' 2.1

JO2 trace 1 39 trace trace

InO trace trace trace trace

101.05 99 .15 1

1 Vogt, Loc. oil.

2 Meyer, "Der Alter Seladon Porzellan," 1885.

3 Ebelman and Salvetat, 3rd Memoir.

4 Vogt, Loc. cit., p. 562.

678 /AMERICAN CERAMIC SOCIETY.

were used. According to the latter, the body was prepared from
an impure dark-burning kaolin softened by the addition oi mi
caceous pegmatite. The glaze was a combination of the pegmatite
with iron-bearing clay and milk of lime, the lime having been
ignited several times with fern fronds. This corresponds with
the description given by Pere d'Entrecelles two hundred year!
ago in the first reports sent to Europe of the ceramic art in China.

These analyses are combined with the analyses of this piece of
very old Chinese pottery in Table 2. The analyses are quoted
from the article by Vogt as the original papers are not availj
able.

From the higher percentages of iron, lime, and alumina, and
lower percentages of alkalies, it would seem probable that a natural
clay and not a modified kaolin had been used in the body of the
Yuan pottery. Ferrous iron is not reported in any of the other
analyses, probably because no attempt was made to determine
the state of oxidation of the iron present. Otherwise the com-
position of the two oldest glazes are quite similar, and do not
differ greatly from the most modern. It is quite interesting that
without the aid of accurate methods of analysis, in fact, depend-
ing in their selection of material largely on visual inspection, the
Chinese potters have controlled the composition of their glazes
within very narrow limits over a period of probably ten centuries.

University of Michigan,
Ann Arbor, Mich.,
December 14. 1918.

THE HYDRAULIC PROPERTIES OF THE CALCIUM
ALUMINATES.'

By P. H. Bates, Pittsburgh, Pa.

Introduction.

The Pittsburgh Branch of the Bureau of Standards has been eon-
ieting several investigations dealing with the hydraulic proper-
es of all of the compounds in the lime-alumina-siliea system,
hese have been carried on in connection with a general in-
stigation of portland cement. It will be recalled that the
)mpounds of this system are :

CaO.SiOj 3CaO.Al.O3 Al,03.Si(J, Ca0.Al2O3.2Si02

3Ca0.2Si02 5Ca0.3Al203 2CaO.Al2O3.SiO2

2CaO.SiOs CaO.AUOj 3CaO.Al2O3.SiO2

3Ca0.Si02 3Ca0.5Al20;i

Dme of these exist in several forms, notably the dicalcium sili-
tte or orthosilicate, which can exist in four — depending upon the
lanner in which it is melted or the temperature from which it is
)oled.2

Such an investigation should be of considerable interest and
lould furthermore be carried on in connection with a similar
ae dealing with portland cement, because, while it does
>ntain materials other than lime, alumina or silica, these
irm ninety per cent or more of this important structural
laterial. At the same time, the limits within which its com-
osition may vary cover but a small part of the total area of the
:rnary system. This restricted composition of portland cement

the result of evolution, that is, it has been found by years of
ial that a cement of about the average composition of commercial
rands can be more economically made and gives better results

1 By permission of the Director, Bureau of .Standards. Read at the
ecember, 1918 meeting of the Pittsburgh Section of the American Ceramic
oeiety.

2 Rankin, "The Ternary System Lime-Alumina Silica." Am. J. Science,
), Jan., 1915.

680 JOURNAL OF THE

in application than others having materially different composi-
tions. Thus the natural cements, containing frequently very high
amounts of magnesia or widely varying amounts of siliea or
alumina, could be produced more cheaply but require Longed
periods of hardening in order to develop strength equivalent
to that of the portland cements. Under certain conditions sla^s
also develop slowly but ultimately very satisfactory hydraulic
properties. But even the area embraced by the compositions
of these three materials (portland and natural cements and
slags) covers but a small part of the area of the system. There]
fore, the entire system offers fields for exploration to see if any
other portions of it contain compounds which under the action
of water possess the property of hardening either in air or water,
and whether after such hardening they remain either weather-
resisting or are able to resist the further action of water. These
two properties, especially the latter, constitute "hydraulic proper-
ties."

While it has not been possible to produce all of the compounds
of the system in the desired amounts and purity, the entire system
has been covered by burnings on batches having compositions
very similar to those of all of the different compounds. It was
found that the following alone developed hydraulic properties:
2CaO.Si02 (/3-form) ; 3CaO.Si02; and all of the aluminates except
the tricalcium aluminate. The fact that the triealcium aluminate,
the only aluminate present in portland cement of normal com-
position and burning, does not have the desired hydraulic proper-
ties is very striking. At the same time one is impressed just as
much by the fact that the other aluminates of lime all have
this property of hardening and acquiring at early periods strengths
in excess of those acquired by portland cement at the end of an
equal period. The aluminates are, therefore, worthy of especial
consideration and the remainder of this paper will be given over
to a brief discussion of their production, their properties, and
the chemistry of their hardening.

3CaO.Al203.

The tricalcium aluminate (3CaO.Al203) can be prepared in
a pure condition with some difficulty. It decomposes at about

AMERICAN CERAMIC SOCIETY. 681

350 C into CaO and liquid. If this decomposed material is
ground and reburned at a temperature of about 13500 C for
^eral hours, it is possible to secure the aluminate free from any
1O. This ground pure tricalcium aluminate, on the addition of
iter, agglomerates into masses, the exteriors of which alone
s hydrated. At the same time there is a vigorous evolution
heat -the mass giving off steam copiously. By grinding
a mortar it is possible to secure a fairly smooth paste free from a
2at amount of the hard agglomerated granules of the material
lich have not been acted upon.
If pats of this paste are placed in a damp closet, they seem

acquire both an initial set (in about three hours) and a final
: (in from 48 to 72 hours). But if the interiors of the pats
2 examined they will be found to be quite soft, and frequently,
they are allowed to remain in a damp closet for about a week,
ey become very soft throughout. If placed in water they
integrate completely. The product of hydration is hydrated
calcium aluminate, which is at first in a colloidal (gelatinous)
■m but is rapidly converted into a crystalline form.
The question naturally arises — does this tricalcium aluminate
t in a similar manner in normal portland cement? If so, what
isks or retards this action — which apparently does not take place
len cement is gaged with water? If it does not, what is there
the cement which so modifies the properties of the tricalcium
iminate so that it appears to have hydraulic properties?
Before considering these questions, it is well to remember
at alumina (A1203) is not required in portland cement — except to
ike it possible to produce it on a commercial scale. A cement
ving all of the properties of portland cement and containing
is than one-half of one per cent of alumina can be produced,
cement of this kind, however, cannot be produced in a rotary
n or at a temperature which would make its production a
mmercially feasible proposition. In the manufacture of
rtland cement the presence of alumina is necessary to the
vering of the temperature of burning to such a degree that a
'nary system may be formed (without too great an expenditure

heat) and from which the tricalcium silicate — the active,
rdening constituent — may separate.

682 JOURNAL OF THE

1 1 a section of portland cement clinker is examined it will
be noticed that the constituents do not have a well defined outline
nor are they separated into well defined crystals. The dicalcium

silicate the major constituent exists in fairly large grains which
blend off into a more or less homogeneous boundary containing th<
tricalcium aluminate and tricalcium silicate as a bonding ma-
terial. Although in a well-burned clinker having a high silicg
content the tricalcium silicate does exist in well-defined crystals
it should be remembered that such a clinker is an exceptiona
one. It can be readily seen, therefore, that in ground clinkei
the grains will not be separate grains of any one constituent
as they would be in a mechanical mixture, but that even the fmesl
grains possibly contain some of all of the constituents. Con
sequently, the activity of the tricalcium aluminate will be re
tarded to a degree by the inability of the water to reach all of it;
surfaces. Although what would be its normal activity in the pun
condition is thus reduced to certain degree, the other aetiv<
constituent of the cement — the tricalcium silicate — is being actec
upon by the water and the lime is going into solution. This lattei
very materially modifies the activity of the tricalcium aluminate
If this aluminate in the pure condition is mixed with hydratec
lime, in amounts not exceeding 10 per cent, and be then gagec
with water, a very smooth, plastic paste, which hardens in air quit<
similarly to portland cement, is obtained. This action is strik
ingly different from that of water on the tricalcium aluminat<
when hydrated lime is not present. It will be recalled thai
in the latter case a smooth paste was not obtainable, nor on<
which would harden — on account of the almost instantaneou-
action of the water and the production of a granular paste com
posed of grains of unhydrated material protected on their ex
teriors by a film of hydrated material.

It appears, therefore, that in portland cement the tricalciun
aluminate does react to a degree in a manner similar to its actior
in the pure condition, but that this activity is reduced to a limitec
degree by the inability of the water to reach all of its surface
and is masked by the presence of a preponderance of the mon
slowly acting silicates. The action is also very materially modi
fied by the presence of lime. The latter may have been eithei

AMERICAN CERAMIC SOCIETY. 683

inally present in the cement in the uncombined condition or
d by the water from the triealcium silicate.
1 view of the fact that the triealcium aluminate, in the pure
iition, does not have any hydraulic properties and cannot be
lily gaged with water, no test specimens were made for de-
lining its tensile or compressive strength.

5Ca0.3Al203.

here is no great difficulty in preparing 5CaO-3Al203. This
riinate when ground reacts so energetically with water that it
lot be readily handled. Some heat is also evolved, but not
he amounts evolved by the triealcium aluminate. In order
etard somewhat the rapidity of the hydration and thus to
ait the molding of the gaged material, it is necessary to use
excess of water. This results in low strengths. If about
r cent of ground gypsum is mixed with it, a smaller amount
ater is required. The strength in this latter case compares
favorably with that developed by portland cement after the
i period of ageing. The products of the hydration of this
linate are hydrated triealcium aluminate and hydrated al-
ia, in other words, during the hydration, alumina splits
ind separates as gelatinous hydrated alumina. On account
3 rapid set, little further work was pursued with it. Several
1 burns were made of it although there were no attempts
icure a quantity sufficient for the preparation of neat, sand,
mcrete test specimens.

CaO.Al203 and 3CaO.5Al.2O3.

irnings of small quantities of CaO.Al203 and 3Ca0.5Al203
ed that both of these set slowly and harden rapidly and
n such high strengths that it was deemed advisable to make
s in a rotary kiln and in sufficient amounts to make concrete
mens for long-time tests. It was also thought advisable
itermine the effect of such impurities as silica and iron oxide
le mineralogical composition and on the physical and chemical
2rties. Eight burns were made, the analyses of the products
lich were as follows:

684

l< HJRNAL OF THE

70 1 80
jCsA

. — The location of the burns (Table i) in the ternary system; also t
fields showing the final products of crystallization according to Rankin.

Table i.

Burn No. 1 2

Per Per

cent. cent.

Si02 0.44 0.76

AI2O3 62 .31 6l .25

Fe203 0.51 o . 60

CaO 36.69 36.32

MgO 0.36 0.48

Ig. loss 0.07 0.50

3

Per

cent.

55

4
Per

cent.

IO.48

46.71

2 .13

39-79
1 .04

5
Per
cent.

6
Per
cent.

17.23 17.38
39.96 3052

2-57 I 85
38.84 46 72

I .29 2.27

7
Per
cent.

II -33
47.06

3IO
34 87

3-66

0.08 0.32 0.14 0.78 0.17

100.38 99.91 100.30 100.47 100.03 99.52 100.19 i°°-

Temperature of

burning ° C. . .

14600

14800 14900 13800 1455 °

13600 1445° 1500

Constituents1....

3C5A

CA CA 2CAS 2CAS

2CS 2CAS CA

CA

5C3A 3C5A CA CA

5C3A CA 3C5.

2CS

2CS 2CAS 2CS 2CS

CA 2CS 2CA

1 Note C =

CaO.

A = A1203. S = Si02.

AMERICAN CERAMIC SOCIETY.

685

. 2. — Some of the burns produced clinker rich in crystals of 2CaO.A\o( ), ■..-

S1O2 and 3Ca0.5Al203. These cannot be distinguished in thin

sections; usually the binary compound predominated.

This figure is a photomicrograph of a thin section

of clinker from Burn No. 5. Mag. oox.

is of more than passing interest to note that this work was
;d and the petrographie examination of the clinker was
: by Mr. A. A. Klein, former petrographer of the Bureau of
lards, before the paper by Rankin2 of the Geophysical
ratory, giving the fields of stability of all of the compounds
s system, appeared. We were very gratified upon noting
is report that the compounds found in these clinkers by
; were those that should occur from their composition —
ding to the results of Rankin (see Fig. 1).
ere was no particular difficulty in making any of these
> with the exception of Nos. 4 and 6. Both of these batches
ow melting points and consequently had a marked tendency
rm "logs" or "balls" in the kiln. In the case of No. 6,
y slight increase in the temperature produced a slag. On
Loc. cit.

686

I' 11 RNAL l »!•' THE

Fig. 4. — Photomicrograph of a thin section of the clinker from
Burn No. 3. Mag. 50X.

AMERICAN CERAMIC SOCIETY.

687

Fig. 5. — Photomicrograph of a thin section of the clinker from
Burn No. 7. Mag. 140X.

unt of this low fusion temperature, the /3-form of the ortho-
ite was not stable and the clinker had a tendency to crumble
ust. This is caused by the inversion of the /3-form of the
ite to the 7-form — which is accompanied by an increase
Dlume of about ten per cent. The material from burn No.
umbled into pieces which were about a quarter of an inch
iameter, while the material from No. 6 was reduced almost
Dletely to a very fine powder. If pieces of the clinker of
latter burn were removed from the hotter parts of the kiln
cooled quickly, the dusting was entirely prevented. The
r mixes formed at the temperatures show very satisfactory
:er, the first two (Nos. 1 and 2) and the last ones (Nos.
d 8) producing white clinker similar in appearance to that
uced in the burning of white portland cement. The structure
me of the clinkers is shown in Figs. 2, 3, 4 and 5.
le times of set of the cements from the above burns as de-
lined by the use of the Gilmore and Vicat needles were as
ws:

688 I1 'l 'KNAI. ( >!•' THE

Table ->

I mi i Pinal

Hum Gilmoi ■ \ i. ii Gilnwi • Vice)

No Hours Minutis Hours Moulds Horn Mniiil., Hours Mil

i 0 55 o 25 4 45 2 io

<> t" o 5 3 50 o 50

3 6 25 5 15 7 5

4 ') 35 O IO T, 30 I 15

5 ° 4" (> 15 4 45 3 15

6 O 7 0 2 O 4.5 0 [2

7 6 o 3 45 * * *

8 7 o * * * * *

* = set between 8 and 24 hours.

The time of set, when studied in connection with composition
and constitution, reveals the fact that the mixes low in lime and
high in alumina are those which set more slowly. The cement
from burn No. 8, which approaches the nearest in composition
to 3Ca0.5Al203, required the longest time for setting, while,
as previously noted, 5Ca0.3Al203 sets so rapidly that it cannot
be easily handled. The cements from burns Nos. 1 and 2, ap-
proaching closely the compostion CaO.Al203, have what would
be considered a quick initial set but a normal final set. They lie
between the 5Ca0.3Al203 and the 3Ca0.5Al203 in their setting
properties. A comparison of the cements from burns Nos. 5
and 6 or from Nos. 4 and 7, which have closely the same silica
content, shows that the time of set is dependent upon the lime
— alumina content, the cement with the higher alumina content
having the slower set.

In these cements the constituents other than the aluminates dc
not appear to materially affect any of the properties. The ortho
silicate, when present in any quantities, was in the 7-form which
does not hydrate and consequently does not set or acquire strength
The same is also true of the ternary compound 2CaO.Al203.SiO«
Consequently, these must be considered as non-hydraulic diluents

Effect of the Addition of Plaster of Paris. — When the in;
vestigation was first started it was decided to add 3 per cent 0
plaster of paris to half of the material from each burn in orde
to regulate the time of set. Such additions to portland cemen
usually retard the set but when added to the cements in thi

AMERICAN CERAMIC SOCIETY. 689

^estigation it invariably accelerated the set to a marked de-
te. This is evident by a comparison of the following times
set (Table 3) of the plastered materials with the times of set
sviously shown (Table 2) :

Table 3.
Effect of the Addition of Plaster of Paris on the Time of vStt.
Initial set. Final set.

lrn Gilmore. Vicat. Gilmore. Vicat.

Jo. Hours. Minutes. Hours. Minutes Hours. Minutes. Hours. Minutes

I O 55 O 15 4 55 2 35

2 O 30 O IO 2 O O 50

3 • o 25 o 20 6 20 3 25

4. o 10 o 3 1 40 o 25

5 o 15 o 5 3 o o 30

6 flash flash

7 o 35 o 20 4 30 2 30

8 1 5 o 20 4 o 2 45

While the acceleration is more pronounced in the case of the
ments high in alumina, all of the work of the Bureau of Stand-
ds seems to show that, unless used in excessive amounts, plaster

paris when used alone accelerates the time of set. It would
•pear necessary that either lime, or a compound that can set
ne free in the presence of water, must be present in a cement

order that plaster of paris may act as a retarder. This ap-
ies to portland cement as well as to the ones under discussion.
Strength Tests. — For determining the cementing value of these
uminates the following test pieces were prepared:

1. Neat tensile test pieces.

2. 1:3 standard sand mortar tensile test pieces.

3. 1-3 standard sand mortar 2 -inch cube compression test
eces.

4. 6" X 12" cylindrical compression test pieces prepared from
1.5:4.5 and 1:3:9 gravel concretes.

These were broken at periods as shown in Tables 4, 5 and 6,
le results given being the averages of determinations on 3 test
eces. All of the tension and the 2-inch compression specimens,
ith the exception of half of those broken at 28-day and 26-week
sriods, were stored under water. These were stored for one-
uarter of the period in water and in air for the remainder of

<><„ |

loi'RNAl. OF THE

M .£

185

0

cd

a.'

0 0 00 O "-. c c o
". -+ -1- -t - - *o -t

f«- <"l - >0 "N O O 0

ir, -f vC -t ", CS CJ l~

0
in

on'
0

0,'

iO io >o >o C IO Ifl IO
<N -f *C IO "+ — O "t
Ml ■+ >C IO "". <*J ""- "t"

B

0>OOt^toOOC
ror^OO <N \C NO CO
iO fO -t "t - - "0 fO

u

a,

i

u-> O O O io >o 'o C

oooo'*;ot^vo

iO >0 O "tj- io nO io O

io O >o O Q »o »n O
cNOr^00-t-0CO
IO CO io io fO N tO "0

Q

o
on

cd

ni

lOOOlOMlOO>0

VO 00 5> IO N ^ N O
* tJ- to r^ h h to «

cu

O io O >o >o io O O
-1-0*0 h n co io r--

^- IO IO ^ N d rOC—

>>

Q

00
CM

C/3

Oh'

0 O io»oO io >o O

MN0O"N00HfO

r^ioio^cocs io>o

si

0 io>o0 ioO >oO

►h & -<t- >-i r^o o oo

CO IO CO CO i-i ** M IN

CO

cj

CU

0 IOI010IO>00 O
r^ro^O rO fO ^ !0 CO

0/

cd

ioOioioOOOO
rocs CS to 1^ 00 O *J-

■t tO IO tO H (N <N IO

CS

0

u
3
0

X

At

pj

O >oO »o io >o •o O

\OOOr/5,0"-<NC^
rO to to rO - M <N to

toio>oO O >o <o 0

0 CO co oo r^-r^OO

Tj- IO tO N -i — CM \0

0 io >o »o >o io O O
oo c> O *o u-> ^ »o r*»

tONHtOHHfJtO

PL,'

00 ioO >o>oO >o 0
r— r^r^« >o <n tJ- to

totin^»N'0'0

81

■a

V

H

a

■o
u

0

7.

Burn No. i

Burn No. 2

Burn No. 3

Burn No. 4

Burn No. 5

Burn No. 6

Burn No. 7

Burn No. 8

-a £
o £

0

—
—

■^T

+j

CS

T3

I-.
CU

O

—

CO

S

cu

—
c«

l_

<u

0

C

CD

w II «

(J cu ^

AMERICAN CERAMIC SOCIETY

691

S.

4;

r>

Vh

01

m

3

~

O*

H

O

l-c

Z

cu

w

u<

X

m

-

T)

01

d

El

w

O

>

a*

w

Li

ed

u

U

u

s

Oh

iOOioioQioOC
10 00 — 'O ci 00 "h vc
ccioOtor^oOci
ci to to f) h. r, -j-

Ph'

10 O "~. 0 0 C >o 10

■tO-O-^CH't
C -f- CO CI Tf CO -N C
f, ^ ^1 O N CN lO 00

to
U

IN

X
'J

Oh'

ioOOOOioO'O
■^-t^O c< ^ 10 <0 l>
Q\r}--H Q\ 10 O tJ-tJ-

-t-»ot^Tt-tocoior^

C0000»o0>0
vC 10 <*0 t-» C t^- 00 C

OOOOvOdfNTfO^O
PO t|- "C}- (N <N m to "+

■A
6

(X

toOOO'OioO'O

1- O 10 00 Tj-r)-(r>-+
OcnO"+00coOO

vo "0"h to ^ + >o 00

01

00000>o>oO
co r^ vc 1^. O O C oc
^ 10 r. •t 0 x »0 co
"Orj-r^cOfN « co^O

CO

>.

Q

0

0

Li

Ph"

1000>0>OIOO>0

►hcOi-hOvOcncOtJ-

0 O fN t)- tj- 0 vo r-~

1 + CO IN H CI -+

•0 >o O >oO O >o 0

rj- co 00 \D r~ 00 0>O
CKr^>-H 10 0 *C (N (N
co-rj-r-Tt-tN — 1000

CO

>.

0!
Q
00

03

&;

io>oOO>oOO<o
"t 10 M 10 ft to r-~

HI

O 10 »i ifl O >o >o O

t> to »i 'O >o >-i h co

r^io^O O" c0"+^O
tNtOt)-(N«tNHHC0

Oh'

ioioOO>oO>oO

"HClOOl^cOOOO

05 \o 0 too » 0 c

\0 r- 00 vO (N tor^O

u

cfl

p3

10 in 10 10 O to O <o

lOtNOONfltO't
N t> 't CN CO "t O 00

to to \o io»-h ** tor^

>>

CD

Q

u
cu

CD

p5

Ph

>oOOOOOOO
tOOtoOOO rt-(N O
00 "+ CI CH C7>I^-(N

fl + IN M h H >tf

Oh

P5

10OOOOOOO
■+ vo •+ 00 0 vO tf- >-
t> r- O OO Tj-tN m \o
•O •* "4- ^ « " 1000

CO

u

3
0

a

CM

Oh

ocoooooo

vC totooc^O^O r^f
ooO^D^OOOoor^

N«5„^MMHft

Oh'

0000<oO>00
r^vo -*t- * to^vOvO
100 CI -3-0 cOp-hOO
to + n^HHNft

rt

■a

a
•0

cu
u.

0

w ci co *+ 10 vc r~ 00

00000000

33333333

33333333

mmmmmpamm

692

JOURNAL « '!• THE

>

cu

2250

24SO

1055

1080
2450

eu

1 530
1525

600

655
1470

*i

1275

870

810

1815

4445

(N

© o o © o

■ IO N N IO fO

■ 00 1 ^ f) ^

nC

00©0©«D©0
^ NO NO X — • *D © <N
ID rD O <N — © fO IN
lDfN*DO<N(N"DX

vO

© © ID ID © © © ©

- -C ^ C C i" '' ^'
f. i^ — ^ i- *t O i*
c-. — -f "i — — *■*• no

o
rs

a.'

1580

1385

5 70
600
1700

jj

© © c c ©

• © ID 1^ X —

■ X ID *f ID O

— CD

*o

4180
1760
4525
2565
1270
1515
4450
7610

Q
o

1:12 P.

1580

1280

700
800
1760

o

a

o

u

ro id O O O O
__ • I- CT> O O X

■ ec to i* ^ ^

a

'/I
1-
0.

•3
o

3
0
0,

0©©©©00©

lowOfC^^C't'o

— tJ« On G\ © ("N O ©
* IN 1 N - — ■* r»

>>
Q

X

O U-. O ID ©
• C* • — • — ■ ■* ■*
■ X ■ ■* ■ •)• O N

(N

1000
j 365
305
1500
3415

VO

0©©©OOiDO

fO nj- ^, n - tj- r^

>>

Q

Ph

ID ID © »D ©

■ CT> ■ C] ■ © f- X

■ U-) ■ -<J" ■ Tf C ^
— M — —

2!

ID »D »D ID ©

■ © <N t>. © r-»

■ O (N IN Tt- ^1
— —CM

>©

3270
4150
2175
3570
885
765
4135
6010

u

3
0
X

l:12P.i

1500

1630

315

375
470

1:12

880
Soft
275
455
890

2825
3000
Soft
3145

840
1475
2930

■a
a

03

c
o

o

0.
0
u

Burn No. 1 **.
Burn No. 2.. .
Burn No. 3. . .
Burn No. 4.. .
Burn No. 5.. .
Burn No. 6.. .
Burn No. 7. . .
Burn No. 8. . .

u

K

0

D

o

ii

0,

H

!A

£

V

a

■*

-

,0

0J

3

H

=

*■

V

d
u

41

s

0

=:

0

^

#

&.

« -n = -

AMERICAN CERAMIC SOCIETY. 693

;he period. The concrete specimens, with the exception of some
)f those tested at the end of a year, were stored in a damp closet.
When it was noticed that some of the concrete specimens were
ihowing a retrogression in strength after ninety days' storage
n the damp closet, it was thought advisable to determine the
jffect of drying out on the strength of the specimens. A set
)f test pieces was therefore removed from the damp closet and
;tored in the air of the laboratory until the end of a year when they
Vere broken. In practically all cases, similar test specimens
vere made of both the plastered and unplastered cements. Where
his was not the case, it is shown in the tables.

Neat Cements. — The results of the tests of the neat specimens
ire not given. These were eliminated on account of the fact that
it no period does there seem to be any relation between the
itrength developed by a cement as shown by a neat tension test
)iece and the strength which it may develop in a mortar or con-
:rete. Such has also been found to be the case with portland
:ement and consequently the present revised specifications
)f both the Government and the American Society for Testing
Materials for the latter no longer recognize or require a neat
est piece. The following examples (Table 7) showing the strength
)f certain test pieces made from the cements containing no plaster,
ndicated in pounds per square inch at the end of one year, will
llustrate this:

Table 7.

Burn No. 2 4 6 7 8

Neat tension 145 960 445 700 130

1 : 3 Mortar tension 425 435 295 390 700

1 :3 Mortar compression 443Q 3200 2310 5225 8945

i :6 Concrete compression 3220 2965 1440 4650 6720

A satisfactory explanation of this behavior is not at hand,
some have attempted to explain it by a nice distinction (?)
jetween the "adhesive" and the "cohesive" strength developed
)y a cement, assuming that one with high "adhesive" strength
jives a high testing mortar or concrete whereas one with a high
'cohesive" strength gives a high testing neat specimen. How-
ever, it is rather difficult to appreciate such a distinction.

The examination of the data of these tables shows as the

694 J( >URNAL OF THE

most striking feature the high strength at early periods of all
of the mortars and concretes with the exception of those tnad<
from cements Nos. 3 and 5. At the presenl time, Standard
Specifications for portland cement have no requirements for

strength specimens other than the tensile mortar briquettes.
These require that the 1 : 3 Standard Ottawa sand briquette
should have a strength of at least 200 pounds per square inch
at the end of 7 days and at least 300 pounds per square inch at
the end of 28 days. In Table 4 it is shown that the materials
from 4 of the 8 burns develop a greater strength in 24 hours
than is required of portland cement in 28 days. The strengths
developed by the other kinds of test pieces used are relatively
high when compared with the same kind of specimens made from
portland cement. Generally, cements Nos. i, 2, 4, 7 and 8
develop as great a strength in any form of test piece in 24 hours
as portland cement will develop in 28 days.

Effect of Plaster. — Those specimens made from the cements
to which plaster had been added and which had been stored in
water continuously did not usually develop as high strength
as the specimens made from the cements from which the plaster
had been omitted. This is more pronounced in the case of the
compressive than in the tensile specimens.

Effect of Storage. — The storing of the specimens in the air
of the laboratory — after storage in water for about one-quarter
of the entire period — produced very marked increases in strength
over that developed by the specimens stored in water for the en-
tire period. A 50 per cent increase was common and in some
cases it exceeded 100 per cent. In passing, it might be
mentioned that German specifications for portland cement re-
quire that the 2 8 -day test pieces be stored for the first day in the
damp closet, then for 6 days under water and the remainder of
the time in the air of the laboratory. It is claimed that such
storage follows more closely the manner in which mortar or con-
crete would be handled in practice. However, seasonal or even
daily changes in the atmospheric humidity render such storage
in the air open to serious criticism. These changes materially
reduce the possibility of reproducing results at different times or
places from the same cement.

AMERICAN CERAMIC SOCIETY. 695

Generally, the concrete specimens showed a decrease in strength
ith age when stored in the damp closet. This is not surprising
hen the nature of the hydration of these cements is remembered,
his consists in the splitting off of the alumina from the aluminates
id its separation as gelatinous hydrated alumina and the for-
ation of hydrated tricaleium aluminate. The latter, until
te periods when it tends to change to a crystalline form, is also
ilatinous. All of these gelatinous or colloidal products of hy-
■ation which form the cementing materials act as any other glue,
i the presence of too great amounts of moisture the latter is
>sorbed and in the presence of small amounts it is given up.
'hether water is absorbed or given up by the colloid depends
Don the amounts in the colloid and in the gaseous phase. When
ments, such as those under consideration, form large amounts

colloid, it follows that the changes caused by moisture will
■ very pronounced. With the taking up of large amounts

moisture the strength of the specimens will be materially re-
lced. When the colloid dries out — provided the drying is
)t carried too far — there is an increase in strength. Just as
e moisture content of the atmosphere does not materially affect
ie strength of an organic glue (such as fish or bone glue) so
so it does not decidedly change the strength of mineral glue
lless it contains too great amounts of water.
There was not a sufficiently large number of burns in the
ries to permit of much generalization in regard to the relations
:tween mineral composition or constitution and strength.

is quite evident, however, that the cements with the higher
umina contents develop the higher strengths. The strength

decreased by increasing the silica content — especially if the
ica replaces the alumina. However, when silica must be in-
oduced, much better results are obtained by keeping the alumina
>ntent constant and replacing lime. The one essential in the
imposition is to keep the content of alumina such that the cont-
rition will approach that of 3Ca0.5Al203.
High Strength Cements. — Some of the strengths developed by
e cements were so high that they are worthy of special cita-
!>n. Cement No. 8 gave a 1 : 3 mortar which developed a tensile
i'ength of 770 pounds per square inch at ninety days and 960

AMERICAN CERAMIC SOCIETY.

pounds per square inch at twenty-six weeks (by (In- combine
storage), and a compressive strength of 8,610 pounds at 7 day
and 10,690 pounds at twenty-eight days (by the combined storage
and a 1:6 concrete with a strength of 6,010 pounds at 7 day
and 8,220 pounds at one year (by the combined storage). Cemen
from burn No. 4 gave a 1 :6 concrete having a strength of 3,14,
pounds in twenty-four hours. The average twenty -four-hou
strengths of the 1 : 6 concretes of cements from burns Nos. I, 2, J
and 8 was 3,000 pounds. It should be borne in mind that th
above figures are the averages of determinations on 3 test piece
and consequently some individual test pieces gave higher strengths

CONCLUSION.

It appears from this investigation that it is possible to make
by a method and equipment differing in no manner from that use<
in making portland cement, cements giving in twenty-four hour
strengths as high as those usually developed by portland cement
in twenty-eight days — either as a mortar or concrete. This quick
hardening cement is not quick-setting. It differs from portlan*
cement in being a calcium aluminate high in alumina (preferabl;
55 to 75 per cent). The presence of silica or iron oxide is not de
sirable but they may be present in amounts up to 15 per cent-
provided they replace the lime and not the alumina.

The commercial possibilities of such a cement are not ver
promising on account of the lack of a widely distributed suppl
of alumina. While the actual cost of manufacture would nc
exceed that of portland cement, the cost of the alumina alor
would so raise the ultimate cost of the cement that the use of tr
finished product for any except very special purposes would t
precluded.

This opportunity is taken to acknowledge the valuable assi
tance rendered by Messrs. Klein, Peck and Tucker at varioi
times while the investigation was under way.

)TE ON CERTAIN CHARACTERISTICS OF PORCELAIN.1

By A. V. Bi.KiNiNcKK, Pittsburgh, Pa.

'orcelains vary widely in their physical properties • such as
dulus of elasticity, coefficient of thermal expansion, heat con-
tivity and electrical resistance at different temperatures. It
tot uncommon for instance, to find differences in the modulus
elasticity as far apart as 6,600,000 and 1,200,000. Again, the
fficient of thermal expansion is subject to decided variations
different porcelains as well as for the same material at different
iperatures. With reference to the electrical conductivity, the
Lies observed for different temperatures are subject to equally
ie variations.

Ahese facts are not at all surprising and are explained by the
ious stages of development represented by different porce-
is. In practically all of them we have conditions far removed
n such equilibria as we find in glasses depending upon the
iposition, the fineness of grinding and intimacy of blending,

rate and the ultimate temperature of firing, and finally the
I of cooling. As a result, the amounts of undissolved quartz
cristobalite, of glassy matrix, of undecomposed clay and of
manite fluctuate within wide variations. It is obviously of
at importance whether the quartz content is high or low,
2ther it is comparatively coarse or fine, whether the percentage
feldspar is small or large or whether additional or different
:es are introduced, and last but not least, what the temperature
itment has been. It is only through more refined and exact
asurements that we may obtain a clear notion of the relation
ween the physical properties of porcelain and the several fac-
s governing them.

n the low -fired porcelains — those fired to about cone 10 — it is
dent that the amount of undissolved quartz is comparatively
*e and that of sillimanite small. It is to be expected, therefore,,

1 By permission of the Director, Bureau of Standards.

698 JOURNAL OF THE

thai this product is subject, upon heating, to the changes in sj
cific volume coincident with the transformations of quartz a
cristobalite. It is not unlikely that some of this type of porcela
when made into heavy pieces, is in a state of stress due to
heterogeneous structure which has not been modified by hi
temperature treatment. It is to be expected, on the other liar
that the .structure would gradually approach a more balance
because more homogeneous, state, as higher firing temperatun
which cause far-reaching solution of the quartz and practical
complete dissociation of the clay substance into sillimanite, a
employed. The development of the latter mineral is desirat
owing to its stability, being subject to no transformation chanJ
its density, 3.031, and its low coefficient of thermal expansion.

The undesirable characteristics of quartz in heavy porcelai
are well understood by the manufacturers of sanitary ware wl
avoid its introduction as much as possible and insist upon t
use of French flint — a partially amorphous modification of sili
which inverts into cristobalite much more quickly and probata
is also dissolved more readily by the fused feldspar. It is n
unlikely that the objectionable property of "dunting" and i
minor manifestations do not appear until the amount of quai
exceeds a certain value which has not yet been established,
the case of electrical porcelain subject to high tensions, it is mc
than likely that the permissible content of free quartz remaini
in the porcelain is very much lower if maximum stability
desired. In advocating a high clay content for chemical por<
lain subject to severe duty, E. T. Montgomery, therefore,1 pi
ceeds on sound principles and his conclusions deserve carel
attention.

In the Pittsburgh laboratory of the Bureau of Standards
large series of porcelains have been made in which quartz t
been replaced both by clay and by synthetic sillimanite. In be
cases porcelains of excellent stability have been obtained as sho1
by the resistance of the bodies to sudden heating and coolii
Particularly was this the case by the use of the sillimanite whi
was prepared by grinding together 258 parts of kaolin, 102
anhydrous alumina and 7.2 of boric acid, and heating the m
1 Trans. Am. Ceram. Soc, 18, 88-92 (1916).

AMERICAN CERAMIC SOCIETY. 699

to the softening temperature of cone 20. The calcined ma-
il consisted largely of sillimanite with a small amount of
>mbined alumina. It was again ground and introduced in
body, replacing the flint. As a type, these porcelains, ex-
ally that in which 40 per cent of sillimanite calcine was used
ther with 40 per cent of clay and 20 per cent of flux, when

to cone 16, not only showed excellent resistance to sudden
ing and cooling but also high mechanical strength. At the
2 time the sillimanite increased the firing range of feldspar
elain decidedly. By raising the amount of sillimanite to even
tr proportions, porcelains of still more marked stability could
)roduced. Similar results were obtained through the intro-
ion of other minerals not subject to molecular transformations,
. as fused alumina, zirconium oxide, etc. The thermal ex-
don is decreased with the replacement of quartz by such
tituents. The body obtained with the use of the zirconia
led to be especially desirable for purposes where high mechan-
strength was necessary.

has been known for some time that the dielectric strength of
elains is greatly reduced upon heating. Messrs. Henderson
Weimer1 found that when heated from 240 to 275 ° C, a body
[ to cone 9 showed a reduction in the puncturing voltage down
/30 of the initial value. The same experience holds true also
the electrical resistance of vitrified bodies. In fact, the con-
tivity of such materials increases very rapidly with temper-
e — according to the law of compound interest at the rate of
it 2 per cent per degree C. That this introduces an important
or in the case of porcelains exposed to higher temperatures,
1 as spark plugs, high tension insulators used in Cottrell proc-
installations, etc., is obvious.

1 order to secure information concerning the effect of compo-
>n upon these phenomena, a large number of porcelains were
le and tested. The specimen used for this purpose was a

of 60 mm. outside diameter, 65 mm. height and having a
kness of wall of 2.5 mm. The cup was placed in a suitable
trically heated furnace. Electrical contact was provided by

1 Trans. Am. Ceram. Soc, 13, 469-475 (191 1); also Weimer and
n, Ibid., 14, 280-292 (1912).

700 J1 'I RNAL OF THE

the use of molten solder on the inside of the cup and its immera
in a shallow metal bath on the outside.
Alternating current, 60 cycle, of 500 volts, was employe!

since direct current gave rise to disturbing polarizing efTec
The current passing through the porcelain was measured by mea
of a very sensitive milliammeter or a dynamometer wattinet
The temperature was determined by means of two insulat
thermocouples immersed in the molten solder of the inside of t
cup.

The results of the resistance measurements were plotted as t
logio R against the temperature, in degrees C. The plots w«
approximately straight lines expressed by the relations

log10 R = a — bt,
where R = resistance of the specimen, in ohms,

/ = temperature in degrees C, and a and b are constan

By computing a resistivity factor, based on the dimensions
the cup, it was possible to calculate the resistance per cubic a
timeter of the porcelain. The criterion used in comparing t
merits of each body was the temperature Te at which the mater
has the definite resistance of one megohm per centimeter cube.

The electrical measurements and computations were made
the electrical laboratory of the Bureau of Standards at Washii
ton.

A characteristic curve for a feldspar porcelain is shown in 1
diagram of Fig. 1, for which the value Te = 335° C. Up
studying the resistance curves of a number of porcelains it \
soon apparent that the feldspar was the important factor
lowering the electrical resistance at these temperatures. Inv
iably, the higher the feldspar content the lower was the resistai
and the value Te. By eliminating the feldspar and carrying
maturing temperature to a high point, say cones 18-20, bodies
good resistance at the temperatures involved were obtain
The same result was reached when the feldspar was replaced
synthetic silicates containing beryllium oxide, magnesia or ot jj
alkaline earths. Through the elimination of the feldspar it 1
found possible to reach a Te value as high as 800 ° C, whidl
quite satisfactory when one considers that fused quartz gave H
result of 880 ° C.

AMERICAN CERAMIC SOCIETY.

701

%

k \
\ \
\ \
\ \

\ \
\\

\ \
\\

\ \

\\

V

\

\\

\

\\
\ \

\ \
\ \
\\
\ \
\\
\\

\
\

200 300 400 S00 600

Fie. 1.

also appeared that with higher feldspar content the polar-
m phenomena became more prominent, indicating clearly
this constituent of porcelain plays a role of an electrolyte,
evident, therefore, that for conditions where a high tension
lain insulator is exposed to elevated temperatures it is neces-

702 AMERICAN CERAMIC SOCIETY.

sary to reduce the content of the feldspar and to replace it b
some other flux.

For porcelains intended to be used under severe electrical an
heat conditions, it would appear desirable, therefore, to eliminat
both the quartz and the feldspar as constituents of the body i
the manner suggested above.

ENAMELS FOR CAST IRON.1

By Homer F. Stai.f.y, Washington, D. C.

I. TIN ENAMEL COMPOSITIONS.
Basis of Classification.

The tin enamel compositions for cast iron in common use
ve, for the most part, been developed in factory practice through
lg years of patient cut-and-try experiments by men who had
tie knowledge of chemistry and less of physics to guide them,
recent years a few ceramic engineers have entered this field
d have materially reduced the cost of the enamels in use. This
luction in cost, however, has consisted mainly in substituting
eap forms of chemicals for more expensive forms and the
mination of a few expensive ingredients of little value. The
leral types of enamels had been too thoroughly worked out in
ictice to permit any radical innovations. The fact that these
pes have stood the test of time and are really a "survival of the
test" is in itself proof that they must conform, at least roughly,
the laws of physics and chemistry applicable to enamel com-
sitions.

Many factors determine the availability of an enamel, such as
iibility, luster, coefficient of expansion, strength, elasticity,
or, and hardness and resistance to chemical agents. Enamel
^positions might be classified according to the relation of com-
sitions to any one of these factors, but the most obvious rela-
n is that of composition to luster, as dependent on the ability
the enamel to remain glassy and not devitrify during working
d cooling. Of course, all of the other factors have a modifying
ect in determining the actual compositions used. In fact, any
amel composition is a compromise.

As stated under the discussion of luster,2 boric oxide and lead
ide are the only fluxing oxides that can be present in enamels

1 By permission of the Director, Bureau of Standards.

2 This Journal, i, Sept. (1918).

-oi J( ii'RXA], OF THE

in large amounts without causing devitrification. The amounl
of the other fluxing oxides that may be added without causing
devitrification is so small that it is practically impossible to make
a satisfactory enamel for cast iron unless the sum of these twq
oxides is above a certain minimum. Boric oxide is always used,
and tin-bearing enamel compositions fall into three general types:
( i ) Leadless enamels, (2) low lead enamels, (3) high lead enamels
(Table 1). Of course it is understood that the following composi-
tions (Table 1) are for powdered enamels to be applied to casl
iron:

The Leadless Type. The leadless type of enamel has been
developed in Germany. It is characterized by a very high boric
oxide content. Since the enamel is so high in boric oxide, it
would have a very low coefficient of expansion and would therefore
tend to chip, other factors, such as strength and elasticity, being
equal, if this were not corrected in the rest of the composition.
Therefore, the German enamels high in boric oxide are high in
sodium oxide and cryolite, both of which have a very decided
tendency to raise the coefficient of expansion of enamels. Since
boric oxide has a favorable effect on the strength and elasticity
of enamels, the unfavorable effect of a large amount of sodium
oxide on these properties is counter-balanced. On account
of the solvent action of glasses high in sodium oxide on tin oxide,
the amount of the latter substance is high, compared to American
practice.

Enamel 1 (Table 1) published by Dr. Julius Grimwald1 who has
written extensively on German enameling practice, is a formula
of this type. Clay is seldom used in making enamel melts in
America, and the amount of tin oxide in Griinwald's formula
would be considered excessive. A formula such as Enamel 2
conforms more nearly to American ideas of what a leadless enamel
high in boric oxide should be.2

The Low Lead Type. — Since lead oxide is present in con-
siderable amounts in the low lead type of composition, the boric
oxide can be lower than in the previous type without danger of

1 Stahl u. Eisen, 30, 1204.

2 H. F. Staley and G. P. Fisher, Trans. Am. Ceram. Soc, 15, 626 (1913)-

AMERICAN CKRAMIC SOCIETY.

705

evitrification taking place. The lead oxide generally runs under
3 per cent and the boric oxide over 8 per cent. In American
ractice, considerable zinc oxide is generally used in this type of
lamel. Enamel 3 is an old formula for this type of enamel and
namel 4 is a modern derivative of the same formula.

The High Lead Type. — In the high lead type of enamel the
ad oxide runs from 16 to 25 per cent of the melted weight of the
lamel. Since the lead oxide is so high the boric oxide can be
>w (around 6 per cent of the melted weight), as in Enamel 5,
ithout danger of devitrification. However, on account of
)nsiderations of strength and elasticity, the boric oxide is often
spt as high as 8 per cent in high lead enamels as in Enamels 6
id 7. The amounts of the other fluxing oxides vary widely,
nt any one seldom exceeds 10 per cent. This is the type of
Dwdered enamel for cast iron most commonly used in this country
id many formulas might be given, but Enamels 5, 6 and 7
fable 1) are typical.

Table 1. — Tin Enamel Formulas.
Raw batch for 1000 pounds melted.

itash feldspar

lay

>dium nitrate

)da ash

arax

ryolite

luorspar

mmonium carbonate. .
iagnesium carbonate. .

in oxide

oric acid

nc oxide

uartz

atassium nitrate

atassium carbonate. . .

ed lead

)dium silico fluoride. ..

[anganese dioxide

rsenic oxide

arium carbonate

No. l.

340

55

5 -5
21

555
1 10

5 -5
8

5 5
188.6

No. 2
410

30

240
120

No. 3.

No. 4.

No. 5.

No. 6.

No. 7

380

400

38o

38o

38o

30

25

30

30

30

90

SO

30

30

225

225

165

215

215

IO

40

30

85

85

120

50

50

83

125

92

to8

15
60

30
82
40

94

108

60

94 175

75

90

105

168

90

50

255

1294. 1 1228 1 16.5 1 166 1 135 1 148 1 150

•( )()

JOURNAL OF THE

Tin E
Percentage

No. 1. No.

Silica 24 .46

Alumina 8-44

Potassium oxide 5 .66

Sodium oxide 10.41

Boric oxide 20.33

Aluminium fluoride 4.40

Sodium fluoride 6 . 60

Magnesium oxide 0.26

Tin oxide 18 .86

Zinc oxide

Lead oxide

Silicon fluoride

Calcium fluoride 0.55

Manganese oxide

Arsenic oxide

Barium oxide

NAMKI.S

Composition.

2. No. 1. No. 4.
26.08
7 00
11.22
5.38
8.23
O.4O
2 .36

2.S.88

7 • 36
6.76

9-95
8.23

1 .60

2 .40

9 .20
10.80
8.00
2 .24
8.50
o. 10
o .60

9.40

10 .80

9 .20

No. S.

24 58
7 .00
6 .42
6.50
6 .04

8.50

6 .00

17 .10

No. 6.

2458

7 ,00

No. T_
24.58

I .80

9 .00 9
10.50 5
16.50 25

8.50 12 .00 5.00 5

5.83 3-i

<>o
42
.}<
87

OO'

00

99.97 100. o 100.00 100.08 99.97 100.06 100. 06-

Tin Enamels.
Empirical Chemical Formulas.

No. 1. No. 2. No. 3. No. 4. No. 5. No. 6. No. 7>

K20 o. 1 88 o. 187 0.230 o. 130 o. 130 o. 140 o. 153.

Na20 0.773 0.423 0.222 0.345 0.201 0.294 0.230

CaO 0.021 ... 0.212 o .200 0.295 0.132 o. 145.

MgO 0.018

BaO 0.073 0.052 0.059.

ZnO 0.390 0.258 0.244 0.142 0.269 o. 140

PbO 0.076 0.081 0.159 0.153 0.273.

MnO o .002 I

AI2O3 0.332 0.260 0.138 0.150 0.132 0.157 0.156

S1O2 1 .255 1 . 118 0.882 0.790 0.785 0.847 0.920

B2O3 0.890 0.766 0.229 0.216 0.155 0.231 0.253

SnC>2 0.383 0.140 0.1 18 0.1 14 0.107 0.124 0.135

F.. 0.519 0.436 o .356 0.266 0.295 0.218 0.144.

II ANTIMONY ENAMEL COMPOSITIONS.1

Sodium Metantimonate Enamel Compositions.

Sodium metantimonate has been employed extensively in
this country during late years as an opacifying agent for enamels
1 For detailed treatment of color phenomena, etc., see Staley, Trans,
Am. Ceram. Soc, 17, 173-189 (1915).

AMERICAN CERAMIC SOCIETY. 707

>r cast iron. The compositions used may be divided into three
/pes: the leadless, the low lead, and the medium lead (Table 2).

The Leadless Type. — In Germany leadless enamels in which
)dium metantimonate is the opacifier have been derived from
le leadless tin enamels, similar to Enamel 1 (Table 1), very high
1 boric oxide and containing only one or two other fluxes. In
lis country, however, leadless enamels containing sodium
letantimonate have been derived from the complicated formulas
l use for tin enamels, by the substitution of other fluxes for lead
dde. This has resulted in complicated formulas, rather high in
oric oxide and generally high in zinc oxide. Boric oxide is
icreased as lead oxide is decreased because, when used in large
nounts, it not only does not itself crystallize but it makes the
lamel more viscous and thus prevents the crystallization of
ther compounds that may have been used to partially replace
ad oxide.

The zinc oxide is increased as the lead oxide is decreased in order
) maintain the strength and elasticity of the enamel. By using

complicated formula the amount of each of the other fluxing
rides is kept low, so that tendency for their compounds to
ystallize and cause dull luster is lessened. Enamel 8 (Table 2)

a typical composition.

The Low Lead Type. — The low lead type of composition
)ntains 3 to 5 per cent of lead oxide; the boric oxide and zinc
ddes are usually a little lower than in the leadless type. The
mount of each of the other fluxing oxides is kept low. Enamel
is a typical example. In this the percentage of sodium oxide is
pproaching the danger limit as far as luster of the enamel is
Micerned. Its use would be safer if 50 pounds of borax were
lbstituted for the 45 pounds of soda ash. This would give boric
side and sodium oxide contents like those in Enamel 8.

The Medium Lead Type. — In the medium lead type of com-
osition, the lead oxide runs between 5 and 12 per cent. The
oric oxide is slightly lower than in the previous type, but still
igher than in high lead tin enamels, and the percentage of each

7o8 JOURNAL OF THE

of the other fluxing oxides is kepi Low. Enamels ><> and 1 1 are
typical compositions. It will be noted that in Enamel 1 1 no
calcium carbonate is used, but that a larger amount of fluorspar
is used than in the other compositions. Thus the calcium content
is maintained at the proper point to give good color and opacity.

Oxide of Antimony Enamel Compositions.

The essential difference between sodium metantimonate enamel
compositions and those in which oxide of antimony is used as the
chief opacifier is that the latter contain much larger amounts of
sodium nitrate. Corresponding reduction is made in other
sodium compounds so that the total sodium oxide content remains
about the same. This larger nitrate content may mean that the
oxide of antimony is converted into sodium metantimonate
(Na-2O.Sb.2O5) or simply into antimony pentoxide (Sb205). In-
cidentally the oxide of antimony enamels which the writer has
found in actual use in factories contain considerable cryolite,
and therefore the percentage amount of oxide of antimony can
be lower than in the case of sodium antimonate compositions
containing little or no cryolite.

Table 2.— Antimony Enamel Formulas.

Raw batch for 1000 pounds melted.

No. No. No. No. No. No. No. No

8. 9. 10. 11. 12. 13. 14. 15.

Potash feldspar 390 39° 39° 375 4ID 4*0 410 4°°

Sodium nitrate 35 35 35 25 70 70 90 75

Borax 350 300 250 275 350 300 240 230

Barium carbonate 80 80 70 100 80 75 80 80

Zinc oxide 140 no 85 70 140 100 100 92

Calcium carbonate 30 30 30 .. 30 30 30 25

Fluorspar 73 73 • ■ 100 60 73 45 55

vSodium antimonate 120 120 120 no

Soda ash 45 55 25 .. 25 30 20

Red lead 31 85 102 46 82 117

Calcium fluoride 73

Cryolite 4° 4° 45 45

Antimony oxide 60 60 64 58

1218 1214 1193 1182 1240 1229 1216 1197

AMERICAN CERAMIC SOCIETY. 709

Antimony Enamels.

Calculated percentage composition, melted.

No. No. No. No. No.

8. 9. 10. 11. 12.

Silica 25 .23 25 .23 25 .23 24.26 26.53

Alumina 7.18 7.1S 7.18 6.90 7.54

Potassium oxide. . . 6.59 6.59 6.59 6.34 6.93

Sodium oxide 8.94 10.75 10.52 8.57 8.22

Boric oxide 12.81 11.00 9.15 10.06 12.81

Barium oxide 6.21 6.21 5.44 7.77 6.20

Zinc oxide 14.00 11.00 8.50 7.00 14.00

Calcium oxide 1 .68 1 .68 1 .68 . . 1 .68

Antimony oxide .. . 10.00 10.00 10.00 9.24 6.00

Lead oxide 3 .00 8.40 10.00 ...

Calcium fluoride. . . 7. 30 7.30 7.30 10.00 6.00

Sodium fluoride .. 2.40

Aluminium fluoride . . 1 .60

No.
13.

No.

14.

No.
15.

26.53

26.53

25.88

7-54

7 54

7.36

6-93

6.93

6 . 76

8.86

8.91

7.62

1 1 00

8.78

8.42

5.83

6.21

6.21

10 .00

10.00

9 .20

1 .68

1 .68

1 .40

6 .00

6 .40

5 80

4 50

8.00

11 .50

7 30

4 50

5 50

2 .40

2 .70

2 .70

1 .60

1 .80

1 .80

99.94 99.94 9999 100.14 99 91 100.17 99 98 100.15

Antimony Enamels.

Empirical chemical formulas.

No. No. No. No. No. No. No. No.

8. 9. 10. 11. 12. 13. 14. 15.

K20 0.127 0.125 0.128 0.1 29 0.1 32 0.1 33 0.135 °I33

Na20 o .263 0.311 0.310 o .267 0.288 0.310 0.328 0.293

CaO 0.225 0.223 0.228 o .246 o . 194 o .227 o . 163 o. 180

BaO 0.072 0.072 o .066 o . 100 0.073 0.069 °-°74 ° °75

ZnO 0.313 0.244 °-i93 0.166 0.313 0.224 0.228 0.214

PbO 0.025 o °75 o 092 • • ■ 0.037 o 072 o. 105

AI2O3 0.1 29 0.127 0.130 0.131 0.1 50 0.15 1 0.157 0.158

Si02 0.761 0.754 ° 111 ° 111 0.801 0.803 0.818 o .808

B2O3 0.332 0.282 0.240 0.277 0.332 0.287 0.231 0.225

SbaOs 0.063 0.063 o 064 o .062 0.038 o .038 0.041 o .038

bj o. 170 o . 169 o . 173 o .246 o .238 o .269 o . 188 0.216

These compositions can be divided into the same three types
as sodium antimonate enamels, the characteristics of each type
being practically the same as before. Enamel 12 (Table 2) is
typical of the leadless type, Enamel 13 of the low lead type and
Enamels 14 and 15 of the medium lead type.2

2 Compare R. E- Brown, Trans. Am. Ceram. Soc, 14, 740-755; H. F.
Staley, Ibid., 17, 173-189.

A METHOD FOR THE DETERMINATION OF AIR IN
PLASTIC CLAY.

Hy II. Spurrier.

Although almost universally used in the pottery industry
in the preparation of plastic clay bodies, the potters pug-mill
is not an efficient piece of apparatus. Though a certain amount of
pugging is necessary, an over-liberal treatment may injure an
otherwise good body by the introduction of an excessive amount
of air. In experimenting with pug-mills it was found desirable
to determine the actual air content of various samples of clay.
In making this determination it was evident that a method which
actually separated and measured the contained air would be
the most likely to give satisfaction. With this in view, some
experiments were made in order to determine the feasibility of
disintegrating samples of pugged clay in boiling water. So
encouraging were the results that it was determined to extend
the preliminary work so as to include the measurement of the
contained air in extruded samples, thrown samples, and also in
sections cut from clay filter press leaves. These simple experi-
ments afforded some quite unexpected information in regard
to the structure of the blanks examined.

The manner in which a piece of clay slakes down in water in-
dicates at once the method of its formation. In an extruded blank,
made on a special machine, we observed that the separation of the
clay particles took place along longitudinal lines — indicating
that the air cavities lie along parallel planes (Fig. i, a). Thrown
blanks exhibited very plainly the spiral striae formed in the
process of throwing (Fig. i, b). A. section cut from a filter-press
leaf did not slake or disintegrate in boiling water but remained
indefinitely in the water without slaking down (Fig. i, c). This
latter case shows clearly that it is the disruptive effect of the gas
vesicles, expanding under increasing temperature, that causes the
•clay to slake down, thus allowing the gases to escape.

AMERICAN CERAMIC SOCIETY.

711

Fig. i . — Showing the behavior of clay blanks in slaking down in water.

It became evident that if the air separation could be made
uantitatively it might be made the basis of a clean-cut de-
grmination. Accordingly, an apparatus was assembled to ac-
omplish this end. It was ascertained that kerosene did not have
ny tendency to slake the clay or in any way disturb it, neither
id it dissolve the air.

A heavy 600 cc. flask (Fig. 2, a) was fitted with a two-hole rubber
topper. The larger hole, being evenly cupped at the smaller
rid of the stopper, was fitted with a 7 mm. glass tube just reach-
lg to the bottom of the cup. The smaller hole was fitted with

right-angle bend, one limb extending to the bottom of the flask,
tie other terminating at a conveniently short distance from the
end; this tube was susceptible of connection with either a kero-
2ne reservoir or a supply of boiling water (Fig. 2, b). The larger
nbe was surmounted by a two-hole rubber stopper entering the
nlargement of the measuring tube; the other hole of the stopper
ras fitted with a glass tube that effected connection with and
eached to the bottom of a 2500 cc. bottle (Fig. 2, c) containing

7' 2

Ji >URNAL OF THE

Fig. 2. — Apparatus used in the determination of the air content in clays.

500 cc. of kerosene and connected with a jet-vacuum pump
The whole was mounted on a universal stand as clearly shown
The determination is made as follows :

The duly measured (weighed if desired) blank is placed in th
flask filled with kerosene which is then tightly corked. Kerosen
is allowed to flow from the elevated reservoir until it overflow

AMERICAN CERAMIC SOCIETY. 713

lto the 2500 cc. suction bottle. The whole is then turned through
n angle of about 100 degrees and more kerosene run in so as to
11 the measuring tube and completely expel every bubble of air,
'he apparatus is then placed in erect position and clamped and the
erosene tube connected with the boiling water in the elevated
ter beaker in such a manner that no air enters the tube. This
lanipulation requires no more dexterity than is called for in an
rdinary gas analysis. The boiling water is now run into the
ask, completely displacing the kerosene — the excess water
inning out of the system into the suction bottle — and the screw-
Dck closed. The disintegration of the clay commences at once
ith the consequent liberation of the contained gases, which
se to the top of the measuring tube — duly calibrated in 0.1 cc.
ivisions for a space of 25 cc. and reading from above downward.

water bath is supported around the flask (Fig. 2, a) which is
ccasionally tapped with the finger tips to effect necessary agita-
on. The vacuum pump is set in operation causing considerable
ilatation of the gas bubbles and duly facilitating their collection
i the measuring tube. After about fifteen minutes, the vacuum

broken and upon equalizing the pressure by using the vacuum
ottle as a leveling bottle, as customarily practiced in gas analysis,
le volume of air is read and the temperature taken. The data
i now in hand for calculating the volume of gas to normal tempera-
ire and pressure or any other basis preferred. The following
>sults illustrate the application of the method :

A sample of air-free filter press clay was passed through a pug-
lill and the air content determined to be 9.61 per cent by volume
F the clay sample — calculated at room temperature, no correc-
on for the barometer being necessary. On passing this clay
irough another pug having well-set blades the air content was
)und to be 9.88 per cent by volume.

Another sample was passed five times through a special pug
fter which the air content was found to be 13.18 per cent by
olume.

The method lends itself admirably to the determination of
le proper blade setting of pug-mills and the determination of
le soundness of pugged clay insofar as air content is concerned.

JEFFERY DEWITT Co.,

Detroit, Mich.

;i | I' (URNAL OF THE

DISCUSSION.
Prof. Binns: I have recently had occasion to take up agai

the question of using tin French rolling tabic- instead of (he pJ
mill in the preparation of potters clay. There is no questio:
in my mind hut that the French rolling table is the ideal machin
for preparing clay and its use would overcome a great man
of the difficulties arising from the air in the plastic body. I
does not, however, prepare the clay as rapidly as the pug-mi
and this has been one of the objections to its use.

Mr. Gorton : I would like to ask if a determination of the ai
content was made upon clay which had been wedged by hand?

Mr. Spurrier: No, we do not use the wedging process in ou
plant at present. At one time we used the wedging process L
the preparation of the clay for a special purpose. In this cas
we noticed that the lamination planes in the blanks were paralh
to the axis of throwing and that the distribution of the air wa
very uneven. It has been our experience that wedging is nc
a satisfactory process for the preparation of the clay blanks.

Mr. Barringer: This question of occluded air is, I thinl
particularly important in Mr. Spurrier's line, which I unde
stand is the manufacture of spark plugs, where the eliminatic
of the air is necessary to the production of the maximum m
chanical and dielectric strength. It is not clear to me, howeve
why it is equally important to eliminate the air from other ceram
wares. I can sympathize with manufacturers of white war
etc., who claim that the French rolling table process is too skn
Why is not the pug-mill satisfactory for the preparation of tl
clay for such wares? Would it not be possible to operate a pu
mill under low air pressure and thus gain the desired capacii
together with the complete or almost complete elimination
the air?

Mr. Love joy: There is one device on the market and
patent pending on another which is designed to do that whiij
Mr. Barringer suggests; that is, to shred the clay to release I
air and remove the latter by suction. The inventor who
working on this device claims that laminations and the faulty dr

AMERICAN CERAMIC SOCIETY. 715

which results from laminations are largely due to the in-
ed air.

[r. Spurrier: In regard to the removal of the air by the
loyment of a vacuum, I do not believe it possible to get
good results. When a piece of clay is subjected to a de-
sed pressure, which we improperly call a vacuum, the air
:les expand, it is true, but they are not removed. The air
:le may be dilated but still remains in the clay. It is very
resting to note that if a piece of clay is placed in a bottle
erosene and subjected to alternating increased and reduced
sure, it will behave like a piece of rubber. The air vesicle
)t removed but increases and decreases in size with the varia-
3 in pressure.

.r. Brown: Would it be possible to remove the air from
ly slip in that way?

iR. Spurrier: Yes, that is being done. If you subject a slip
bottle to reduced pressure, it will increase in volume, and if
tap the bottle with the knuckles, a leaden sound is given
L indicating a lack of continuity. If the evacuation has
inued long enough and all of the air is removed, the bottle
ring when tapped. The removal of the air from a heavy
by evacuation is not practicable.

[r. Barringer: The statement confirms my opinion that
nay be worked from a clay by some such method as kneading
is not readily removed by reducing the outside air pressure,
to pugging under reduced air pressure, I believe that this
Id involve considerable mechanical difficulty in the operation
maintenance of the apparatus.

NOTES ON SAGGER CLAYS AND MIXTURES.

My ('.. H. Brown, Xiw Brunswick, N. J.

Introduction.

The sagger problem, although an old one, does not appear
nearer its solution than it was several years ago. The sag
losses in our potteries and other ceramic industries is a seri
one and the life of the saggers is an important item in the pr
and loss columns in most of the plants. It is true that ther
considerable variation, in the length of life of the saggers u
in the different plants, but there appears to be no gener;
accepted practice particularly as to the kinds of clays used in
sagger mixtures. Frequently, a change in the composition
the sagger mixture is necessitated by inability to secure the cl
formerly used and the consequent substitution of new and unt
materials in the sagger batches.

In a great many plants, the composition of the sagger mix
is governed for the most part by the cost of the clays — the qua
of the clays and their suitability for this purpose being seconc
to the price per ton. However, very often, a so-called low-gi
or cheap clay will produce a sagger having longer life than s<
of the so-called high-grade clays sold at fancy prices.

Character of Sagger Failures.

The character of sagger failures may be briefly classifiec
follows :

i. Breakage due to the rough handling of the unburned sag
in placing in the kiln.

2. Breakage due to the too rapid heating of the saggers du
the first burning.

3. Breakage during the cooling in the kilns.

4. Breakage due to rough handling of the saggers in dra1
from the kiln and emptying and in again placing in the kiln, i

AMERICAN CERAMIC SOCIETY. 717

lilures due to the deformation of the loaded saggers at the
am kiln temperatures,
rig up in order the above types of failures :
reaking in Handling. — The breakage in the green state
placing may obviously be reduced by more careful super-
3f the placing of the unburned saggers in the kiln. This
je is, however, governed to a large extent by the strength
sagger clays in the dried condition and clays having good
h should be used — all other properties being satisfactory,
rength in the dry condition is also governed by the size
oportioning of the grog particles and the percentage of
led.

reakage in Heating-up. — Some of the breakage of saggers
e ascribed to their being too rapidly heated during the
irning. This is especially true if the saggers have not
horoughly dried before placing in the kiln. It is also
ed to a certain extent by the amount of rapid heating
the clays used in the sagger mixture will withstand. The
has observed instances in which the facilities for drying
jgers were entirely inadequate and, when placed in the
ley still contained an appreciable percentage of the water
1 tempering the sagger mixture. In starting up the fires
kiln, this water may be too suddenly converted into steam
icking of the saggers results.

must not lose sight of the fact that sagger clays contain
ally combined water, as do all other clays, and that this
led water is only driven off as the kiln reaches a visible
it. If the temperature is raised too rapidly during this period
temically combined water is converted too suddenly into

a strain in the sagger is set up, and cracking and
ge result,
reaking in Cooling. — From the observations we have made,

inclined to believe that most of the breakage of saggers
place during the cooling down of the kilns. Any clay
it will crack if cooled rapidly enough. Some, of course,
ack or break more readily than others. Of the many
b wares, fused quartz has perhaps the greatest resistance

718 JOURNAL OF THE

to temperature changes this being due to its very low coefficieo
of expansion and its very regular and uniform expansion as

contraction when heated or cooled. A piece of fused silica wan
will crack, however, if cooled quickly enough. We find the sami
to be true of all ceramic wares which must necessarily withstan<
unusual temperature changes in use. We might mention clu -mica
porcelain, Marquart porcelain, gas mantle rings, spark plugs
etc. Turning to the heavier and cruder clay products, manu
factured from fire clays, we find the same to be the case. It i
well known that silica brick are very sensitive to temperatun
changes and crack very readily. It is generally accepted that i
fire brick having a high content of quartz or silica will cracl
and spall more readily than the so-called high-clay fire brick
We also know that some high-clay fire brick are more sensitive
to temperature changes than others. Therefore, under the sanu
conditions of cooling, saggers made from one fire clay mixtun
will crack more readily than saggers made from another mixture
We will discuss the reasons for this later on in the paper

4. Breakage from Handling. — Some of the sagger breakage i
undoubtedly due to the more or less rough handling which ttu
saggers receive in placing in and drawing from the kiln. Thi
breakage may be aggravated by the use of a vitrifying wad-cla;
and the consequent necessity of pounding the sagger in order I
loosen it from the one underneath. It may also take place i;
setting the sagger on the floor or on the bench. A sagger ma;
not be actually broken by the handling but it may be so weaken©
that it will crack and break the next time it is fired. Obvioush/
the reduction of breakage from this source is a question of usin
the right kind of wad-clay and of supervision of the drawin
from and placing in the kiln.

5. Deformation. — Failures through the deformation of saggei
— due either to heavy loading or firing to high temperatures-
are not so common as those due to cracking or breaking. Thi
type of failure, of course, is more prevalent in those plants firin
large and heavy wares and in those plants employing the higbf
firing temperatures. Our American pottery practice diffei
markedly from that of Continental Europe in the firing ten

AMERICAN CERAMIC SOCIETY. 719

tures employed — which are uniformly lower in our ceramic
ts. We might say that the average firing temperatures of
dins is not above cone 10 or 11 and very rarely in this country
ire encounter firing temperatures above cone 13 as against
s 16, 17, and 18 reached in Europe. Our sagger problem is,,
ifore, not analogous to theirs. We must repeat, therefore,
deformation is not the most general type of failure of saggers.

Requisite Properties of Saggers.

aving discussed the character of sagger failures we take up
question of the properties requisite in saggers in order to
ce the failures due to the above. The requisite properties
sagger may be outlined as follows:

Refractoriness.

Mechanical strength.

Resistance to deformation.

Resistance to temperature changes.

Refractoriness. — The refractoriness of the clays used in a
er mixture should receive first consideration. Refractoriness
come to mean more than a high melting or softening point,
tie investigations on fire clays during the past ten years have
to a very definite classification of our fire clays, not only in
rdance with their so-called melting points, but also on the
5 of their vitrification or burning behavior at the temperatures
rhich they are to be used. The fact that one fire clay has a
:ning point two or three cones higher than that of another
clay does not necessarily mean that the former will give
er service in a sagger mixture. If a kiln of ware is being
[ to cone 12, it is immaterial whether the saggers used have a
ing or softening point of cone 29 or cone 32. In this case the
ivior of the sagger at cone 13 and below is of infinitely more
artance than is its behavior at cone 26.

igger clays, and all fire clays used as a bond in the manu-
jre of clay products, may be roughly divided into two classes,
i;ar as their burning behavior is concerned, i. e., they may be

ified as vitrifying and open-burning. A vitrifying clay is,
ourse, one which vitrifies at a comparatively low temperature

72o

J( lURNAL OF THE

to

Id

J6

**

$V

$te$

B.— -

*" —

14

-

12

^ 10

6

f

Prtf

2

-IV(

2.1 tT ^

ao

12

V.
8 \

Temperature 'Q

Fig. i.

and retains its vitrified structure and low porosity over a wic
temperature range without swelling or over-burning. In Fig. i
shown the burning behavior of a clay of this type. It will t
noted that the shrinkage of this clay although high is quite un
form and that the trend of the shrinkage curve conforms qui1
closely with that of the porosity curve.

In Fig. 2 is shown the burning behavior of an open-burnii
clay — one which retains a porous structure and high absorptkl
at a comparatively high temperature. The burning shrinka;$
of this type of clay is usually low and shows little change *
firing over a wide temperature range.

In Fig. 3 is shown the burning behavior of a clay which doW
not fall under the above classification. It will be noted that t'l
porosity of this clay decreases quite regularly with the firil

AMERICAN CERAMIC SOCIETY.

721

.X

Sjxrjvf?

.18

/6

1)
/OS

z

0.0,

Temperature "C

Fig. 2.

perature until a vitrified structure is attained at a high tem-
ture. The firing shrinkage increases quite regularly with
firing temperature and the shrinkage curve conforms with the
isity curve, *. e., as the porosity decreases, the shrinkage in-
ses. A clay of this type is obviously not well suited to the
ufacture of many of our clay products as with each increase
^mperature the clay undergoes further shrinkage and conse-
ltly does not reach a condition of equilibrium until a high
Derature is reached.

1 Fig. 4 is illustrated the burning behavior of a clay which
bviously unsuited to the manufacture of saggers or other
products which are burned to high temperatures. It will
toted that the porosity of this clay decreases somewhat uni-
lly with the increase in temperature until it reaches the

722

JOURNAL OF THE

' e rr\p'

Fig. 3.

minimum porosity, after which the porosity increases rapidly,
indicating the development of a vesicular structure.

In view of the wide variations in the burning behavior of the
different fire clays, the softening or fusion point is therefore not
a safe criterion in the selection of a clay to be used in a sagger
mixture.

2. Mechanical Strength. — Under mechanical strength we will
consider the resistance of the saggers to breakage during placing
and drawing after the first firing. The mechanical strength of
a burned sagger in the cold is dependent upon several factors:

1 . The extent to which the bond clays vitrify at the finishing
temperature of the kiln.

2. The mesh of the sieve through which the grog is sieved and
the proportioning of the grog particles.

AMERICAN CERAMIC SOCIETY.

723

0

„ ^

{

1

x&

*-

\

?

<$$>

$*J

\

\

f

, /

f

V

\

\

/ \

\

?

\

\

1

s

9

\

\

>2

\

\

\ ,

1

/"n

0

20

18
16

10

I

5

00

Temperature °C.
Fig. 4.

(. To the presence or absence of hair cracks — small cracks
often developed by sagger mixtures in drying and burning.
L The proportion of grog to raw-clay used and the uniformity
mixing and tempering.

' . Bonding Clays. — A certain amount of vitrifying bond clay
necessary to the production of a sagger of good mechanical
mgth in the cold. By the sole use of an open-burning bond
y the particles of grog are held loosely together and weak
;gers result. By the use of a vitrifying bond-clay the grog
"tides are held firmly together. ' The use of a certain amount
a vitrifying bond-clay having a burning behavior similar to
it shown in Fig. 1 , is desirable in order to secure a sagger of the
ired strength. The greater the amount of vitrifying bond
y used the greater will be the strength of the resultant saggers-

7 >| J( HJRNAL < >F THE

It is not desirable t<> use too much of the vitrifying clay- In
order to secure a sagger of the desired openness a certain amount
of open-burning clay, having a burning behavior similar to that
shown in Fig. 2, must also be used. By confining our mixture
to grog, a vitrifying clay and an open-burning clay, and by vary-
ing the proportions of the two clays, we can secure a sagger of
any desired mechanical strength in the cold. A good proportion
is, say 50 per cent grog, 20 per cent vitrifying clay and 30 per
cent open-burning clay.

2. Sagger Grog. — It has been our experience that too little
attention has been paid to the screening of grog for sagger mixes
and that our knowledge of the best practice is limited. The
work of F. A. Kirkpatrick1 on the effect of the size of grog in fire
clay bodies has thrown considerable light on this subject. In
a sagger mixture in which there is a preponderance of large sized
grog particles, the spaces between the grains are filled with the
bond clay which, in vitrifying, allows the particles of grog to
slide upon each other. Instead of having a skeleton of grog
particles touching at many points and held together by the bond
clay we have a matrix of bond-clay in which are imbedded the grog
particles and the bond-clay becomes the predominating factor
in the behavior of the mixture.

The proportions of the sizes of grog particles secured by screen-
ing through a single sieve are very irregular — this being especially
true if a sieve of large mesh is used. As an illustration, a grog
which was prepared by passing through a 3/8 mesh sieve had
the following proportions of different sized grains :

3/8 - 1/4 mesh. . . .36 per cent

4-8 mesh 30 per cent

8-12 mesh 10 per cent

12-20 mesh 9 per cent

20 - 40 mesh 8 per cent

Through a 40 mesh . . 7 per cent

It will be noted that in this grog the larger particles predominated
and very little shaking caused the finer particles to settle out
leaving a skeleton of coarse grains.

1 Trans. Am. Ceram. Soc, 19, 268-300 (1917).

AMERICAN CERAMIC SOCIETY. 725

j. Hair Cracks. — We have often observed the small hair cracks
in saggers, sometimes present in the green sagger before burning
but more often observed in the burned saggers. These hair
cracks may be caused by the use of a too large-mesh screen in
sieving the grog, to improper proportioning of the grog particles,
to the use of a bond clay which has an excessive drying and burn-
ing shrinkage, or to a combination of the above causes. If there
is a preponderance of large sized grog particles there will be an
excess of bond clay between the grog grains, and the bond clay,
having a greater shrinkage than the grog, decreases in volume
and pulls away from the grog particles causing a strain and pro-
ducing the small cracks which are bound to decrease the mechanical
strength of the sagger. The qualities imparted by coarsely ground
grog in the sagger mixture may be more than offset by improper
proportioning of the grog grains and the resultant unsound
structure produced by the small hair cracks.

The general practice in Europe — employed to a limited extent
in this country — of grading the particles of sagger grog into
different sizes and then adding definite proportions of each size
to the sagger mixture is an excellent one. In this way the voids
between the larger sized grog particles are filled with smaller
sized grog particles and so on until we have a skeleton of grog
particles held together by the bonding clays. A mixture of this
kind will have a smaller drying and burning shrinkage and greater
mechanical strength when burned than a similar sagger mixture
in which there is a preponderance of the larger sized grog particles.

3. Resistance to Deformation. — As we have previously stated,
failures due to deformation of the saggers appear to be fewer than
the losses due to breakage.

The causes of the deformation of fire clay bodies in general
and in particular of fire brick have been very thoroughly studied —
especially by the Bureau of Standards.1 We are all more or less
familiar with the load test as applied to fire brick when heated
to high temperatures. The behavior of a fire brick when weighted
down or loaded is analogous to the behavior of a sagger when
filled with ware. In the case of the fire brick it has been found
that the deformation under a load is largely due to the softening
1 Bur. Standards, Tech. Paper 7.

726 I' WRNAL OF THE

of the bonding clay used in holding together the particles of grog
A vitrifying clay softens and deforms more readily than an "pen
burning clay and hence, if our saggers are failing through bending
or deformation, the remedy is to use a smaller percentage o!
vitrifying clay, or to substitute a refractory open-burning cla>
for some or all of the tight-burning clay used in the sagger mixture

4. Resistance to Temperature Changes. — A large proportior
of sagger breakage appears to take place during the cooling dowi
of the saggers in the kiln. It has been well established that a very
dense vitrified fire clay body has less resistance to temperature
changes than a fire clay body retaining an open structure aftei
burning. A sagger mix prepared from a dense grog and excessive
quantities of vitrifying bond clay is bound to be very sensitive tc
changes in temperature. The structure of the sagger should b<
as open and porous as possible without sacrificing its mechanica
strength in the cold. This, of course, is governed by a propel
proportioning of the contents of open-burning and vitrifying
bond clay used in the mixtures.

The cracking and breaking of saggers is undoubtedly in man)
instances due to the use of open-burning bond clays, having i
high content of sand. The practice of adding sand to a saggei
mixture is not a sound one from the information which is avail
able. The sensitiveness to cracking of all clay products contain
ing a high percentage of silica is well known.

Silica brick, which contain from 93 to 98 per cent silica, an
particularly sensitive to changes in temperature and for thi:
reason are not used in a great many furnaces which are heated uj
and cooled down intermittently. We have encountered a greatei
tendency in the so-called quartzite or siliceous refractories t(
crack and spall than is observed in the so-called high clay re
fractories. We know that red-burning clays, containing a higl
percentage of sand, used in some localities in the manufacture 0
common building brick, are very sensitive to cracking if coolec
down quickly in the kilns. By the introduction into a saggei
mix of an open burning bond clay having an excess of free sand
we are inviting cracking and breaking of the saggers — through th<
inability of the sagger as a whole to withstand the volume change;
of the particles of sand introduced in the bonding clay.

AMERICAN CERAMIC SOCIETY. 727

Flie irregular thermal expansion and contraction of the different
ms of silica is the source of trouble in the manufacture of a
:at many of our ceramic wares. It has given trouble through
j cracking and "dunting" of terra cotta, of vitreous and porcelain
litary ware, fire brick, and in every clay product in which there
an excess of silica, either in the form of sandy fire-clays or in
j form of flint as in our white burning pottery products.
We would, therefore, expect that the introduction of an ex-
isive amount of sandy clay into a sagger mixture is very liable
cause cracking — followed by breaking of the sagger either
the same or in the succeeding burn.

Raw Materials and Composition.

[n view of the above, therefore, it is not sufficient that we
3w the burning behavior of our sagger clays — whether vitrifying
open-burning. We must also know something of their com-
sition, especially as regards their content of coarse or even very
e grained sand or silica. Obviously, the sagger should be as
e from iron as possible and from all granular materials which
: liable to cause popping or spitting-out.

[t is significant that, in the manufacture of ceramic products
ich are particularly designed to withstand temperature changes
i especially of the higher grade wares, it has been found neces-
y to eliminate silica in the form of flint from the body mixtures.
1 may mention particularly the porcelain tubes used in insula-
g the thermocouple wires of pyrometers. The well-known
irquarat1 tubes are known to have a high content of alumina
i little or no free silica was used in the preparation of the
iy. The same is true of the porcelain pyrometer tubes now
ng manufactured in this country. Flint has been eliminated
m the body and calcined aluminous mixtures have been sub-
tuted in its place. Flint has also been found objectionable
the special porcelains which are subjected in use to tempera-
'e changes — such as gas-mantle rings, porcelain insulators,

Attention should be called to the use of kaolin in sagger mix-
"es. In addition to being highly refractory and open burning,
Trans. Am. Ceram. Soc, 18, 268 (1916).

728 JOURNAL OF THE

the kaolin does not contain excessive amounts of sand as do

sandy open-burning fire clays. Very satisfactory refract
bodies have been prepared from mixtures of kaolin, vitrify
fire clay and grog. The mixtures have low drying and burn
shrinkage, and have great resistance to deformation at I
temperatures The use of kaolin, however, is not aeo
providing we can secure a clay which retains a high porosity
the firing temperature. Deposits of secondary kaolin occur
this country, which, although they may not be satisfactory
use in the manufacture of high grade pottery, would be I
valuable in the manufacture of saggers. We have found that
addition of kaolin in amounts as low as 10 per cent materia
improves the refractoriness and other qualities of sagger or otl
mixes.

The composition of the grog used in a sagger mixture is govern
of course, by the composition of the rawr clays used in the mixta
i. e., if the grog is prepared by the crushing of old saggers
addition of a small quantity of kaolin into the mixture tends
maintain the refractoriness of the grog.

Preparation of the Mixtures.

Although most of our sagger mixtures are prepared by the so;
ing pit method, we would call attention to some of the advanta
of using the ordinary wet-pan in mixing and tempering sag
mixtures. We have found that by the use of a wet-pan a mt
more uniform mixing of two or more clays is secured than by
use of a soaking-pit. The soaking-pit method depends upon i
slaking down of the clays in water. If our sagger clays sli
down readily in water the soaking-pit method is not so objecti
able. However, we know that some fire clays will remain
water for a long time without slaking down. Clays of this
should not be used in batches prepared by the soaking-pit meth
The writer has seen saggers, made from a mixture of red and 1
clay and grog mixed by this method, in which the red and yel
particles of the bond clays used could still be distinguished in
molded sagger. The use of the pug-mill in connection with
soaking pit does not always overcome the objections to

AMERICAN CERAMIC SOCIETY. 729

ing-pit method of preparation. The heavy mullers of the
pan crush the particles of clay and thoroughly blend the two
lore clays used in the mixture.

re have also found that by the use of the wet-pan a larger
entage of grog can be introduced into the sagger-mixture
lout the loss of its necessary working qualities. This is
mnted for by the fact that the wet-pan develops the maximum
ticity of the bond clays and hence increases their carrying
icity for grog. The advantage of the increased grog content
he lower shrinkage of the mixtures. Saggers prepared from
pan mixes also mold better and with smoother surfaces,
bjections have been raised to the wet-pan method on the
inds that the heavy mullers of the pan crush the larger parti-
of grog and make it too fine. This objection may be over-
e1 by the use of a wet-pan having mullers which are raised
ltly off the bottom of the pan. By the use of a pan of this
i, the particles of grog slide underneath the mullers without
ig crushed.

Summary.

l summarizing, we may say, therefore, that the length of life
aggers is influenced by a great many factors, not the least
hese being the necessity of care in the preparation of the
tures and in the drying and first burning of the completed
jers.

1 order to produce a sagger giving maximum service, a
■ough knowledge of the composition and burning behavior
he clays used in the sagger mixture is necessary.

Department of Ceramics,
Rutgers College.

1 /. Am. Cerum. Soc, 1, 18 (January, 1918).

MAGNESIA WARE.

By Trvgvb l>. Vbnbbn.

Readers of Mr. Ferguson's recent note in This JourbI
on the Sintering of Magnesia, might get the impression, qu
unintentional on the part of the author, that the world is a
has been absolutely dependent upon Germany for magne
ware. It is true — as far as the writer is aware — that magne
wares, such as crucibles and tubes of American manufactu
have not thus far been sold in the open market in this count
but it has been produced in considerable quantities in met
lurgical laboratories. The process of manufacture — althou
requiring electrical furnaces of considerable capacity — is qu
simple, and has been fully described in the technical literature.

More than ten years ago, Dr. C. F. Burgess, who was th
connected with the University of Wisconsin, obtained a pate
on a process of manufacturing MgO crucibles, consisting in ph
ing crushed or powdered MgO, previously sintered or fused,
a graphite mold, the inside part of which was split so as to forn
heating element through which an electric current could be pass
for the purpose of heating the magnesia to a sintering tempei
ture. It is not known whether this process was ever successfu
applied. The author has tried forming magnesia crucibles wi
out the use of binders by means of graphite molds heated in
electric furnace to 18000 C. While the product was strong a
sound there was strong evidence of reduction of MgO by 1
graphite of the mold, with the result that the crucible was !
uniform and furthermore that the inside of the mold was consuir
to such an extent that it could be used only once. It is suppo:
that similar difficulties were encountered by Dr. Burgess in 1
application of his patented process. However this may
magnesia was fused at the University of Wisconsin and the fu: '
MgO crushed and made into magnesia crucibles before 19:
1 J. B. Ferguson, "Note on the Sintering of Magnesia," J. Am. Cer\
Soc, 1, 439 (1918).

AMERICAN CERAMIC SOCIETY.

73i

» work has been continued since that time by Watts,1 Kowalke
others, and magnesia crucibles have been in common use
re.

ince 191 2 the author has been making magnesia crucibles for
Dratory purposes almost without interruption. A description
he method used was first published in 1914.2

Fig. 1.

lalcined magnesite was fused in an electric arc furnace, the
•d MgO crushed to go through a 40-mesh screen, then mixed
hi about 5 per cent hydrated magnesia, molded into crucibles
1 steel mold, dried and finally heated to 18000 C in an elec-
furnace. The crucibles thus made were uniform, strong, had
e shrinkage and showed no sign of softening at 18000. Chem-
1 O. P. Watts, "Making Magnesia Crucibles," Wisconsin Engineer, 17,
1913)-

* Yensen, "Magnetic and Other Properties of Electrolytic Iron Melted
'aeuo," Eng. Expt. Sta., University of 111., Bull. 72, appendix 1, 49-51
4).

732

JOURNAL OF THE

ical analysis showed the finished product to contain about •>
per (cut MgO, the remainder being chiefly SiO^, lW).) and AU):
A purer product of still higher refractoriness could be obtains
by using a ])urer material to start with, but for the purposes ir
tended by the author the above amounts of impurities caused n
trouble.1

•

i

f8/<?-|

«♦♦♦
I'M

♦ • *

'

Fig. 2.

During the last two years magnesia crucibles of various sizes
been made in considerable numbers at the Westinghouse Rese
Laboratory, a description of which wTas published last ye
A photograph of some of these crucibles is shown in Fig.
The method used is essentially the same as that employed at t
University of Illinois.

1 Results published by the Bureau of Standards show that the melti
point of pure MgO is 28000 C, Bur. Standards, Set. Paper 212, Jui

1913-

2 Yensen, "Preparation of Pure Alloys for Magnetic Purposes," Tra i
Am. Electrochem. Soc, 32, 176 (191 7).

AMERICAN CERAMIC SOCIETY. 733

The author has made attempts to persuade manufacturers
refractories to take up the manufacture of magnesia ware such
crucibles and tubes, but thus far without success. It seems
though the manufacturers are not yet convinced that there is
ufficient market for this kind of ware to warrant going into the
siness, and yet there is no product on the market today that
1 be used as a successful substitute for it as a container for the
rpose of melting pure iron and iron alloys. Until some pro-
ssive manufacturer can be persuaded to hazard the under-
ling, it seems therefore probable that each individual user
1 have to manufacture his own magnesia crucibles.
[t has recently been brought to the author's attention that
rstals of MgO may be of value for certain optical purposes,
rge crystals were separated from pieces of fused MgO and ex-
iined by Dr. P. G. Nutting who found these crystals to be
Dical and isotropic and to have very high refractive index
i low dispersion — properties that indicate valuable lens ma-
ial. This phase of fused magnesia has been described in de-
l by Dr. Nutting at the Baltimore Meeting of the Optical
:iety, December 27th. A photograph of some of the crystals
>hown in Fig. 2.

Research Laboratory,

Westinghouse ElEC. & Mfg. Co.,

East Pittsburgh, Pa.

ACTIVITIES OF THE SOCIETY.
Action by the Board of Trustees.

The motions listed on page 587 of the Journal as pending haf

been passed by the Board without further amendment, exod
in the ease of motions 3 and 5 in the list proposed by the Pres
dent. These have been amended and passed in the followjj
form:

3. "That, as of the date of November 1, 191 8, fifty (50) eopi<
of each volume of the Transactions be reserved for sale in con
plete sets only, and that the Secretary be instructed to purchasi
at five dollars ($5) each, sufficient copies to bring the total stoc
of each volume to that figure. It is the sense of this resolutio
that a complete set shall mean nineteen Volumes, Nos. 1 to i<
inclusive, of the "Transactions of the American Ceramic Society.

5. "That copies of the volumes of Transactions and of tt
Journal shall be furnished to members who have or shall alio
their dues to become delinquent only after the number reserve
for sale has been provided. Furthermore, that in cases in whic
the volume that should be furnished in consideration of a certai
year's dues is not available, the member shall have the optic
of selecting any available volume, or paying only the different
between his annual dues and the price to members of the Yolun
or the subscription price of the Journal to members.

ACQUISITION OF NEW MEMBERS DURING
DECEMBER, 1018.

Associate.

Dr. P. G. H. Boswell, The University, Liverpool, England.
S. Paul Ward, Rancagua, Chile, South America.
J. S. Laird, University of Michigan, Ann Arbor, Michigan.
Harry W. Fenton, Akron .Smoking Pipe Co., Mogadore, Ohio.

AMERICAN CERAMIC SOCIETY. 735

J. Iv. McAllister, J. L. Mott Co., Trenton, N. J.

Frank E. Latter, 2389 Mance St., Montreal, Canada.

J. P. Callaghan, Sharon Clay Products Co., Sharon, Pa.

C. B. Young, Ohio State Brick & Stone Co., Newark, Ohio.

Maurice A. Smith, McKee Glass Co., Jeanette, Pa.

Harry A. Truby, Pittsburgh Plate Glass Co., Creighton, Pa.

H. Homer Knowles, Box 427, East Liverpool, Ohio.

Fen wick D. Foley, St. John, N. B.

Paul R. Morris, 300 East 9th Ave., Tarentum, Pa.

N. B. Radabaugh, 1572 Rydal Mt. Rd., Cleveland, Ohio.

LOCAL SECTION MEETINGS.
NEW ENGLAND SECTION.

Boston, Mass, January 11, 1919.

he meeting was called to order by Chairman A. A. Klein
he Boston City Club, at 7 p.m., with twenty members present.
Ir. Klein spoke of the recent death of Mr. C. L. Walduck
his loss to the Society. It was voted that a committee be
ointed by the chair to draw up resolutions and that copies
sent to his parents and to the American Ceramic Society.

ater in the evening the committee reported as follows :
We, members and appointed representatives of the New
[land Section of the American Ceramic Society, in remembrance
he sterling qualities of Charles L. Walduck personally and his
rient services as Secretary of this Section, wish to record our
;h felt loss of him as a man and as Secretary of this Section,
extend our sympathies to his parents and recommend to our
ent organization, the American Ceramic Society, a suitable
:uary notice in their Journal."

W G. Whitmore,

R. C. Purdy,

M. F. Cunningham.

'he officers elected for the coming year were :
hairman, W. H. Grueby, Grueby Faience & Tile Co., Bos-
Mass. ; Secretary and Treasurer, Charles J. Hudson, Norton
I Worcester, Mass.; Councilor, Charles H. Kerr, American
ieal Co., Southbridge, Mass.

736 JOURNAL OF THE

Mi. R. C. Purdy spoke interestingly on the purpose oi tj

Section both as regards itself and its relation to the America

Ceramic Society. Two interesting and valuable talks followed:
"Some Problems of the Glass Manufacturer," by C I

Kerr of the American Optical Company and "Clav Testir

Methods and Purposes," by Carl II. Lawson of the Xortc

Company, Worcester, Mass. A general discussion of the papl

followed.

During the afternoon previous to the meeting the membl

enjoyed a trip through the chemical laboratories of Arthur I

Little, Inc.

Charles J. Hudson, Secretary.

PITTSBURGH DISTRICT SECTION.

At a meeting held at the Mellon Institute, University of Pitt
burgh, Pittsburgh, Pa., December 19, 1918, the merging <
the former Beaver Section into the Pittsburgh District Sectio
was completed. Morning and afternoon sessions were held.

At the opening of the morning session by-laws were adopt*
and the following officers were elected for the ensuing yea
R. R. Hice, Chairman, C. R. Peregrine, Vice -Chair man, Thos. I
Sant, Treasurer, F. H. Riddle, Secretary and F. W. Walker, Si
Councilor.

After the election of officers a varied and interesting progra
of technical papers was afforded as follows :

"Industrial Research," Dr. E- W. Tillotson, Mellon Institul

"The Effect of Some Fluxes on Porcelain," F. H. Ridd
Bureau of Standards.

"The Influence of Some Ball Clays upon the Casting of CI
Wares," J. W. Wright, Bureau of Standards.

"The Hydraulic Properties of the Calcium Aluminate
P. H. Bates, Bureau of Standards.

"Note on the Electrical. Resistance of Porcelains at Tempe'
tures up to 8oo° C," A. V. Bleininger, Bureau of Standards.

Members desiring to become affiliated with the above Secti
should communicate with the Secretary, F. H. Riddle, Bu«
of Standards, 40th and Butler Sts., Pittsburgh, Pa.

JOURNAL

OF THE

MERICAN CERAMIC SOCIETY

A monthly journal devoted to the arts and sciences related to
silicate industries.

. 1 November, 1918 No. 1 1

EDITORIALS.

GROWTH OF THE SOCIETY.

he American Ceramic Society has been in existence twenty
's. During that time it has grown in membership from 22 to
'ly 1 100. The Society represents industries, the value of
se output — including clay products, glass and cement — aggre-
:s $450,000,000 annually in the United States. The Society
:s to occupy towards the clay, glass and cement industries the
e position of leadership that the American Chemical Society,

American Society of Civil Engineers, the American Society
Mechanical Engineers, the American Institute of Mining
[ineers, and some others, occupy in their respective fields.
: 100 per cent increase in membership of the Society during

past two years indicates the need of an organization of its
i and augurs well for the future welfare and growth of the
iety.

'he Society should, however, profit by the experience and
mple of other technical organizations and endeavor to co-
mate its activities through permanent officials, having no
zr business in life than the advancement of the Society's
rests. With the steady growth in membership and the
ening and intensification of its activities through the Journal,
al Sections, Professional Divisions, etc., the Society's business
lecoming too large and varied to depend upon the part time
"ices of its committees and executive officers. With the in-

738 AMERICAN CERAMIC SOCIETY.

creased revenue through the accession of new members and tin
gratifying support from linns and corporations through conj
tributing memberships, the time is fast approaching when the
Society will be fully able to financially support the coordination

of its activities through full-time executive officers. Let us lend
our efforts in this direction!

KILNS.

There is no problem of more importance facifig the manu
facturer of ceramic products than that of the selection of the typ<
of kiln best suited for the burning of his wares. The low efficienc)
of the average periodic kiln renders it the greatest source oi
economic loss in most of the plants where used. The rising cosl
of fuel and the necessity of its conservation in the future render tht
kiln question an increasingly vital one.

The discussions in recent numbers of this and other Journals
as to the efficiencies and applicability of the available types oi
kilns to the burning of the different kinds of ceramic product?
leave the selection in most cases a wide open one, and a verdici
has not as yet been reached. The question recently raised in th<
editorial columns of a contemporary Journal as to the design anc
efficiency of the modern by-product coke ovens should giv<
ceramic engineers and manufacturers serious food for reflection
Have we arrived at the millennium in kilns or will the futun
produce radical types far superior to any which have been placec
upon the market thus far? It is doubtless true that no one typ<
of kiln will be developed which will be applicable to the burnins
of all grades of ceramic wares.

The suggestion that at the next annual meeting of the Societ]
there be a symposium on kilns should receive attention at once
At that time the presentation of reliable data as to the relativ
efficiencies of the different types of kilns now in use and ataj
unbiased reports as to the service rendered will be a step forwani
in the solution of this vexing problem. In the meantime, ther:
is need for more data as to the efficiencies of present installations
and it is to be hoped that ceramic engineers will continue to collec
and make data of this kind available.

ORIGINAL PAPERS AND DISCUSSIONS.

A CONTRIBUTION TO THE METHODS OF GLASS

ANALYSIS, WITH SPECIAL REFERENCE TO

BORIC ACID AND THE TWO

OXIDES OF ARSENIC.

By E. T. Allen and E. G. Zies.

In the course of the work on optical glass undertaken by the
eophysical Laboratory in the year 191 7-1 8 some investigation
as made of the well-known Jena optical glasses. It included
nong other things complete chemical analyses of a considerable
imber of them, to determine the composition in doubtful cases
id to see if anything could be learned respecting the relation
composition to quality. In this connection the authors spent
msiderable time in testing the accuracy of certain analytical
ethods and in devising others, an account of which is here
ven.

We had supposed that the system of silicate analysis so well
orked out for rocks by Hillebrand and others could be applied
ith little or no change to glasses. But glasses contain certain
sments like boron and arsenic seldom found in rocks in ap-
eciable quantities; many such elements, in fact, if all sorts
glass are considered. Again, glasses contain certain combina-
3ns of elements which are difficult to separate and determine ac-
irately by ordinary methods. On the whole we have found
em considerably more difficult to analyze than rocks.
E. C. Sullivan and W. C. Taylor have realized these difficulties,
id have made valuable contributions to the subject,1 especially
ong the line of rapid analytical methods. Our work may sup-
ement theirs in some measure. We have devoted our attention

1 J. Ind. Eng. Chem., 6, 897 (1914).

740 JOURNAL OF THE

for the most part to optical glasses, and our chief purpose lias bei
to improve the accuracy of the methods. At the same tim
in the work done on boric acid and the two oxides of arscni
we have also kept in mind the more genera] utility of the method!
I. Determination of Trivalent and Pentavalent Arsenic.

It may be confessed at the outset that the time and ipal
devoted in this paper to the determination of arsenic will probabl;
seem entirely disproportionate to the rest of the subject. Th
chief reason for the more detailed study of this determinatioi
was our interest in the role which arsenic plays in glass making
The percentage of arsenic in glasses is so small that if we are t
follow successfully its history from the batch to the finished glas
we must have an unusually accurate method for its determina
tion. When later it became evident from our work that Good
and Browning's very accurate iodometric method for arseni
acid could be applied to the determination of arsenic in all sub
stances where it could first be accurately transformed into th
sulphide, we felt further justified in the course we had followed

Glass commonly, but not invariably, contains arsenic. Al
the Jena optical glasses which we have examined contain il
The total arsenic in glass may be directly determined by fusin
a weighed portion with alkali carbonate and nitrate, removin
any nitric acid which remains by evaporating to dryness wit!
sulphuric acid, filtering off the silica and insoluble sulphate
if any, and precipitating the arsenic as sulphide; this ma
finally be transformed into magnesium pyroarsenate and weighec
or reduced by hydriodic acid and titrated with standard iodin
solution.

When a mixture of arsenic acid and arsenious acid is evaporate
in platinum with hydrofluoric and sulphuric acids, the arseniot
acid is completely volatilized as the trifluoride, while all the ai
senic acid remains in the residue entirely unreduced. Thus th
two oxides are very accurately separated from one anothei
the pentavalent arsenic may be determined in the residue an:
the trivalent arsenic by difference between the pentavalent an
total arsenic. The trivalent arsenic can also be directly determine
in the distillate if the operation is carried out in a gold or platinui
still of suitable form.

AMERICAN CERAMIC SOCIETY. 741

1. Determination of Total Arsenic. — There is no element
originality in this determination; we have simply proved by
ithetic experiments that no arsenic is lost in a fusion with
hum carbonate and niter, and that fairly satisfactory estima-
ns of the total arsenic in glass may be made after separating
: arsenic as sulphide, oxidizing the latter with hydrogen peroxide
1 ammonia, precipitating the arsenic from this solution as mag-
;ium ammonium arsenate and finally weighing it as pyro-
enate. Further, if the solution formed by oxidation with
Irogen dioxide and ammonia is reduced by hydriodic acid and
ated by a standard iodine solution, after the method of Gooch
1 Browning, the determination becomes highly accurate and
y be done in far less time.

Volatilization of Arsenic in the Process of Fusion Involved
Analysis. — Unfortunately for the purposes of analytical con-
l it is not possible to prepare a glass which shall contain a
en amount of arsenic trioxide and probably not possible to
pare one carrying a given amount of arsenic pentoxide. The
sible loss of arsenic by volatilization during the fusion with
ium carbonate which is involved in the decomposition of a
>s for analysis, had therefore to be tested indirectly. To
; end we took mixtures of arsenious oxide in known quantity
I a powdered glass containing no arsenic. To these were added
. pure dry sodium carbonate per gram of glass, the fusion was
de as in an ordinary determination of silica and the arsenic
; determined in the product. An appreciable quantity of
mic was lost.

Taken o .0530 g. AS2O3

Found 0.0797 g- Mg2As207 equivalent to.. 0.0508 g. AS2O3

Loss 0.0022 g. As203

n a second experiment :

Taken 0.0710 g. As2Cs

Found 0.1058 g. Mg2As207 equivalent to... 0.0674 &• As203

Loss o . 0036 g. As203

Vben a small amount of niter, about 0.1 g., was added to the

7 1 2 JOURNAL OF THE

flux a still greater quantity of arsenic trioxide was fused with

loss.

Taken o.ioio g. As2()j

Found o 1389 g. MgtAsjO; equivalent to o 1013 g. As/),

O.OOO3 g. As/).,

The determination of the arsenic was made as outll
above and described in detail further on. Now, although I
quite possible that arsenic which is combined in a glass mi
be entirely retained when the glass was fused with sodium c
bonate alone, it is undoubtedly safer to add niter, for the c
bonate sometimes contains bits of organic matter, possibly pa
or cardboard, which would almost certainly lead to the volatili
tion of metallic arsenic during the fusion of the glass. C
venient details for the method follow :

Fuse 1 g. of powdered glass with 3 g. pure, dry sodium c
bonate and about 0.1 g. potassium nitrate. When cool, set
crucible in a 12 cm. casserole, add water, cover and put in 10
1 : 1 sulphuric acid.1 Warm until the fusion cake is remov
clean and remove the crucible and evaporate till white fui
appear. Heat somewhat longer until the silica becomes de
enough to filter without trouble; cool, add hot water and 1
till any insoluble sulphate settles satisfactorily. Barium gi
trouble here, because in evaporating to dryness the sulph
is partially redissolved in the concentrated sulphuric acid am
again precipitated in a fine state of division by the addition of wa
Filter with suction, wash a few times with boiling water {
suck dry. The filtrate may be kept down to 100 cc. For f
the bulky precipitate may retain a little arsenic, open the fi
and transfer the precipitate to a 20 cc. platinum crucible m
the aid of a stout platinum wire, add a little sulphuric acid an
cc. hydrofluoric acid and evaporate to white fumes. G
transfer the residue to a small beaker with a little hot wa
boil and filter, adding the filtrate to the principal one. To
filtrate from silica add a fragment of potassium iodide, heat ;

1 Hydrochloric acid cannot be used here because, as is well kn(
arsenious acid is thus volatilized as the trichloride, and even arsenic ac
partially reduced and volatilized.

AMERICAN CERAMIC SOCIETY. 743

•ipitate the arsenic with hydrogen sulphide. The precipita-

is materially accelerated by the reduction of the arsenic acid
liydriodie acid. When the precipitate has well coagulated,

partially in order that the solution may absorb some excess
lydrogen sulphide, stopper and set aside. If the solution is
red immediately the results are likely to be too low.
xperiments were made on the precipitation of the arsenic
1 a strong hydrochloric acid solution, without previous separa-

of silica, but so much silica came down with the sulphide
;ing so much subsequent trouble that this method was aban-
sd.

f course if antimony is present it will have to be separated
1 the arsenic by one of the regular methods. If the glass
;ains lead, the precipitate will generally have an orange
r, but the color is due to the presence of a little sulphide of
. If the sulphate solution is allowed to stand before treat-
it with hydrogen sulphide till the lead is completely precipitated
sulphide precipitate is a clear yellow. The addition of arsenic

antimony together has been tried in glass making, but all
glasses we have examined contained either one or the other

not both. After standing overnight, the arsenic sulphide
Itered and washed with hot water acidulated with sulphuric

he sulphide is then treated on the filter with a few cc. of
ng ammonium hydroxide, and the flask rinsed with the same,
filter is washed two or three times with a few cc. of hot water.

shows any residue after the solution of the arsenic it is trans-
id to a 150 cc. beaker and digested warm with ammonia and
tie ammonium carbonate. The orange precipitates mentioned
ye have to be digested a short time before they show the dark
ivn color of lead sulphide, and the suspicion is that till then they

retain arsenic. The soluble portion is added to the principal
nic solution, which is conveniently handled in a 250 cc. beaker,

oxidized by hydrogen peroxide — 3 to 5 cc. of a 3 per cent
tion is sufficient. Cover the beaker and boil for some time
the hot plate to complete the oxidation and decompose the
;ss of peroxide. We have confirmed the statement in the text-
ks that a complete oxidation is accomplished in this way

744

JOURNAL OF THE

Levol's Method1 for the Determination of Arsenic, If tl:
arsenic is to be determined gravimetrically, the solution is qJ
precipitated with magnesia mixture, the precipitate is washJ

dried at 1300 and then slowly heated in the electric oven to 55c
where it appears to be stable. The precipitate gains again on tfi
balance (1 to 2 mg. in the case of a large precipitate), and tr.
final weight agrees best with the quantity of arsenic taken, thoug
this may be due to a compensation of errors. Our results by th
method on pure arsenic solutions were satisfactory, as the sequi
shows, but the results on glasses did not show the accuracy whic
our ultimate purpose required. (See p. 760.) It will be n
membered that one cannot control his results by working with
synthetic glass containing a known quantity of arsenic; he mus
depend on the agreement of duplicate determinations and ti
comparison of different methods. The following table sho*
the agreement of results which one may obtain when the Lev<
method is applied to the determination of arsenic in glasses.

Table 1.
Total arsenic in some Jena optical glasses determined as Mg2As207, showii
the agreement of duplicates. Results in milligrams per gram of glass.

No.

1

2

3

4

5

6

8

9

11

15

Kind of glass.
*Schott's
numbers.

03832

0340

0154

0144

0167

021 1

02071

01209

0722

0599

Mg2As207. ■ • .

Mean

Equivalent to
As206, ,

Mean

Variation in
AsjOs

7.0

5-8

6.4
5-2

4-3
4-8

0.9

5-2

65
5-8

39

4.8

4-3
0.9

2.9
2.7

2.8

2 .1
2 .O

2 O
O. I

3-9
4.0

4.0

2.9
3-0

3.0

0. 1

3-4

3-6

3-5

2.5
2.7

2.6
0 .2

6.1

7-2

6.6

4-5
5-3

4 9

0.8

7.1
8.3

7-7

5-2

6.2

5-7
1 .0

5-5

6.2

5-8

4-i
4.6

4-4

0.5

7-2

8.1

7-7

5-3
6.0

5-7

0.7

3-7
4-1

3-9

2-7
3-0

2.9
0.3

* See Doe
for the approx

1 See espe
Set., [4] 9, 55

Iter's
imate
daily
(1900)

"Hanc
compo
Levol,

Ibuch
sition
Ann.

der M

of som
chim.

ineral
e of tl

phys.,

;hemie
tese gl<

[3] I?

," Bd

isses.

, 501;

. I, 6,

Austi

pp. 8;
ti, Am

1-88
. Jm

AMERICAN CERAMIC SOCIETY.

745

le method may serve very satisfactorily for large quantities
senie, but for these small amounts the percentage error is in-
issibly high. As the Levol method is the handiest method
:he determination of arsenic where only occasional results
required, we spent some time in the effort to increase its
racy.

hen the results in Table i are compared with those obtained
by the Gooch and Browning method (see Table 3) a number
lem are found to be considerably too high. Several mag-
im ammonium arsenate precipitates obtained in the analysis
lasses were examined for insoluble matter like silica, lead
late or barium sulphate, but solution in a few drops of warm
e hydrochloric acid failed to reveal any. When the filter
residue were incinerated, never over 0.2 mg. was found,
le only possible clue to the high results which we have is
abnormally high determination obtained by precipitating
lesium ammonium arsenate in the presence of considerable
ogen peroxide.

Taken.

Found.

S2O6.. . O.OII6 g.

'2O2. . . 5 cc. 3 per cent solution

otal volume about 25 cc.

Mg2As207 o .0392

As205 0.0290

e do not know the meaning of this result, but if hydrogen
ride can cause too high results in the precipitation of arsenic
lagnesia mixture, some of our errors in the determination
senic by the Levol method may have been due to the incom-
removal of hydrogen peroxide from the solution after the
ition of the arsenic.

rther work on this line was abandoned when we had tested
olumetric method described below.

ethod of Gooch and Browning. — Much more accurate than
,evol method is the method of Gooch and Browning1 accord-
:o which the arsenic acid is reduced by hydriodic acid to
ious acid which is then titrated with a standard solution

Am. Jour. Set., 3, 40, 66 (1890).

746 JOURNAL OF THE

of iodine in the usual way Gooch and Browning worked on
with pure solutions of arsenic acid, so that we were obliged
test the applicability of the method to solutions formed by tl
oxidation of arsenic sulphide with hydrogen peroxide and an
monia.

For these and subsequent tests we used a standard solutk
of arsenic acid made as follows: First a sample of arsenioi
oxide was selected from several in stock on the basis of micr
scopic examination. The sample selected was practically hom
geneous showing no intermixture of a higher refracting substan
suggestive of the presence of antimony, as some others did. F
this examination the authors are indebted to Dr. H. E- Merw
whose experience in such work is quite exceptional. Solutk
of a portion of the sample in hydrochloric acid showed no residi
of sulphide. This sample was sublimed at low temperatu
several times, only the first fraction in each sublimation beii
retained. The powdered product was then dried over calciu
chloride. For some work it may be that a more rigorous criterk
of purity would be required, but the experiments which folk
show this preparation to have been quite pure enough for o
purposes.

1.0003 g- °f this preparation of As203 was put into a platinu
dish and covered with a watch-glass. Concentrated nitric ac
was added and evaporated on the water bath, the operati'
being once repeated. Finally the temperature was carried
about 200 ° on the electric hot plate and held there till the furr
of nitric acid could no longer be detected. The arsenic a<
was dissolved and made up to 100 cc.

10 cc. of this solution was precipitated and determined

Mg2As207.

Taken o . 1000 g. As203 equivalent to o . 1 161 g. As2Os

Found o . 1569 g. Mg2As207 equivalent to.. . o . 1 161 g. As205

The solution was prepared and evaluated before the adopti
of the Gooch and Browning method. For some of the te-
another more dilute solution was made in the same way from 1
same sample of arsenic trioxide. Measured volumes of the stai
ard solution of arsenic acid were diluted with wrater and mi'

AMERICAN CERAMIC SOCIETY. 747

dedJy acid, 5 cc. 1 : 1 H2S04 to 100 cc. total volume, a little
issium iodide was added and the arsenic was precipitated by
rogen sulphide. The washing and oxidation of the sulphide
i already been described. When, in the course of the oxida-
, the ammonia has been expelled and the solution has been
iced to 10 or 15 cc. by evaporation on the hot plate, about 7
[ : 1 H2SO4 is added and the boiling continued about fifteen
tites. Some oxidizing agent1 formed by the hydrogen dioxide,
Dtless a mixture of nitrite and nitrate, is thus removed. If
hur or sulphide of arsenic is precipitated when the acid is
id, the oxidation is of course incomplete. The Gooch and
flrning method is carried out advantageously as follows : Pour
oxidized solution into a 300 cc. Erlenmeyer flask marked
. a "glass" pencil at the levels to which 40 cc. and 100 cc.
quid fill it. Dilute the solution to the 100 cc. mark with hot
;r and add a fragment of potassium iodide — 0.3 g. is sufficient
small quantities of arsenic. The two-bulb trap used by
ch and Browning is set into the neck of the flask, and the
ated iodine is expelled by vigorous boiling. A stream of
on dioxide led in through a capillary tube prevents bumping

aids in driving out the iodine which the solution retains
ciously. Gooch and Browning evaporate to 40 cc. and then
)ve the iodine which remains by rapid addition of very
:e sulphurous acid in required quantity. We regard it as a
j more accurate to remove the iodine entirely by repeated
ng. To this end add another portion of water (25 cc), also
her fragment of potassium iodide to insure complete re-
ion, and boil down again to 40 cc, repeating the operation
the solution is colorless. Then dilute quickly with cold
:r to 75 or 100 cc. and complete the cooling by whirling the

in a vessel of ice water. Pour in 12 to 14 cc of a "saturated"
ion of potassium carbonate, which has the advantage over
lm carbonate of greater solubility. This neutralizes the
>r part, but must not neutralize all, of the sulphuric acid.

neutralization is completed by solid sodium bicarbonate.

starch paste and titrate. It is not well to cool the solution

Ilosvay de Nagy Ilosva, Ber., 28, 2031 (1895); Bull. Soc. Paris, 3, 2,
1889).

748

JOURNAL OF THE

below, say, about 200 or 25 ° since the speed of the reaction wit
iodine is then considerably retarded. We use a solution containir
about 1.1 g. iodine per liter of solution, so that 1 g. of solutio
is equivalent to about 0.5 mg. As205. The standard solution

Table 2.

Showing the accuracy of the arsenic determination by the iodometric meth<
when the arsenic is first precipitated as sulphide, oxidized by H2Oj,
reduced, and titrated.

Grams iodine solution re-
quired

Equivalent to mg. As205

Mg. As206 taken

Error

1.

2.

3.

4.

47I1

785'

II .822

23 302

2 . I

3-5

5.8

115

23

3-5

5.8

11 .6

0.2

0.0

0.0

0.1

57 95J
28.6
29 .0

0.4

weighed in a weight burette.3 The very satisfactory resull
in Table 2 show that the Gooch and Browning method for arseni
may be made general, or at any rate that it may be used in a
cases where the arsenic can be previously transformed into tr
sulphide without loss.

The method was now applied to the determination of tot
arsenic in a series of glasses. The process was that described c
page 742, except that the arsenic acid was reduced and titrate
with iodine instead of being weighed as pyroarsenate of maj
nesium. The results appear in Table 3. In these 14 pairs
duplicate determinations, in each of which 1 g. of glass was use
the maximum variation is 0.3 mg. and the mean variation only 0
mg. The method is to be commended for its speed as well :
its accuracy.

1 1 g. of this solution was equivalent to 0.000446 g. AS2O5.

2 1 g. of this solution was equivalent to 0.000494 g. AS2O5.

3 For improvements in the iodometric metric method for arsenious ac
see E. W. Washburn, /. Am. Chem. Soc, 30, 31 (1908).

AMERICAN CERAMIC SOCIETY.

749

Table 3-

atal arsenic in glasses determined iodometrically, in duplicate,
milligrams per gram of glass.1 ■

As206 in

Jo. of

rlass.

Kind of

glass.

\ Jena
IO383:

\ Jena
(O340

\ Jena
/O144

\ Jena
IO167

\ Jena
JO211

\ Jena
\ O2122

jjena

/ O2071

G. iodine

solution

used.

34
04

58
50

24
49

60

05

11 95

11 94
13.20
13-32

Mg.
AS2O5
found.

32
3-2

2.8
2-7

2-9
2.9

3-2

2.9

3-4
3-1

5 3
5-3

5 9
5 9

No. of

glass.

13

15

Kind of
glass.

Jena

O1209

\ Jena

)0722

\ Jena
/0578

\ Jena
/0599

[P. P. G. Co.
1 Spectacle
[ Crown

(P. P. G. Co.
/ Green Plate

P. P. G. Co.
Dense Flint

(".. iodine

solution

used.

I I

48

I I

58

13

74

14

25

IO

82

IO

03

6

10

6

17

21

7i

21

.58

5

.61

5

.62

19

•7i

19

56

Mg.
As2Os
found.

5-1
5.1

6.1
6.3

5-3
50

30

3.0

The Separation of Trivalent from Pentavalent Arsenic2 and
le Determination of Each. — The separation of the two forms
f arsenic can be accurately made by heating the mixture in plat-
lum with hydrofluoric and sulphuric acids. The arsenious acid
; changed into the trifluoride which is easily and completely
olatilized, while all the arsenic acid remains in the residue,
'he operation is conveniently carried out by heating the material
1 a capacious and loosely covered platinum crucible or small
ish by means of an electric hot plate to a final maximum tem-
erature of about 200 °. We used crucibles holding 40 or 60 cc.
nd raised the lid slightly by setting a piece of heavy platinum

1 In determinations 1-11, inclusive, iodine sol. I was used; 1 g. solution
quivalent to 0.000446 g. As208. In determinations 13-22, inclusive, iodine
at. Ill was used; 1 g. solution equivalent to 0.000494 g. As2Os.

2 This separation has not been worked out for mixtures which contain
ntimony.

750 JOURNAL OF THE

wire across the top of the crucible. Bumping was prevented
by dropping a piece of crumpled platinum foil into the crucible

along with the contents and evaporation was accelerated by sur
rounding the crucible by a fire-clay cylinder.

Standard solutions of arsenious and arsenic acids were used
in this series of experiments. The preparation of the standard
solution of arsenic acid has already been described (see p. 746
The arsenious solution was made by dissolving a weighed portion
of As203 in a little warm caustic soda solution, diluting sufficiently
and acidifying with dilute sulphuric acid. The solution was then
cooled and made up to a given volume. 10 cc. arsenious solu-
tion containing o. 1000 g. Ag-iO.j was mixed with 1-2 cc. 1 : 1 H2SO4
and 5 cc. HF and evaporated on the hot plate to white fumes. The
residue was cooled, diluted with water, neutralized by sodium
bicarbonate and tested with starch paste and dilute iodine solution ;
the first drop struck a blue color. The solution was now acidified
with hydrochloric acid, made alkaline with ammonia and tested
with magnesia mixture. No precipitate was obtained.

Again 10 cc. of the same arsenious solution was treated in the
same way. This time the residue was dissolved in water and tested
with hydrogen sulphide. No arsenic was found. All trivalent
arsenic is therefore volatilized by heating a substance containing it
writh hydrofluoric acid in the manner described.

5 cc. of the arsenic acid containing 0.0585 g. As2Oo was now treated
with sulphuric and hydrofluoric acids in a similar manner. The
residue was dissolved in water and precipitated by magnesia
mixture.

Taken o .0585 g. As203

Found 0.0791 g. Mg2As207 equivalent to... . 0.0581 g. As2Oi

Error o .0004 g. As205

This shows that arsenic acid is not reduced or volatilized by
hydrofluoric and sulphuric acids.

When the pure arsenious solution is evaporated on the hot plate
to white fumes (max. temp, about 200 °) with sulphuric acid alone
no arsenic is volatilized.

.

AMERICAN CERAMIC SOCIETY. 751

Taken o . iooo g. AsX ).-,

Found 0.1562 g. Mg2As207 g. equivalent to. 0.0995 g- As2Os

Error o . 0005 g. AS2O6

In Table 4 (page 752), mixtures of the two standard
olutions were used, and to test at the same time the validity
f the method for glasses, 1 g. (or in determinations 6 and 7
or lack of material, 0.5 g.) of powdered glass free from arsenic
fas added to the mixture. In determinations 6 and 7 a flint
lass containing 45 per cent lead oxide was used. These results,
ke those immediately preceding, were done before the method
f Gooch and Browning was adopted; hence the arsenic was
weighed as magnesium pyroarsenate. They show in a convincing
ray that the method suffices for an accurate separation of the
wo oxides of arsenic while the presence of powdered optical glass
1 no way interferes with it. Also the conditions under which the
eterminations are made so nearly simulate those of a determina-
ion of the arsenic in glass that one can hardly question that the
lethod will suffice to determine the two forms of arsenic in a
lass — at least an optical glass where substances which might
ossibly interfere with the process, such as iron and chlorine,
re present in minimum quantity.

S. R. Scholes1 believed he had determined both oxides of arsenic
1 glass, in the following way: A weighed portion of glass was
issolved in hydrofluoric acid and then evaporated with sulphuric
cid till the hydrofluoric acid was driven out. In this solution
ie reduced the arsenic acid by hydriodic acid and titrated the liber-
ted iodine in the acid solution with standard thiosulphate.
Ie then made the solution alkaline with sodium bicarbonate
nd titrated the total arsenic with standard iodine. We now
:now that the solution he used could not have contained any
f the arsenious oxide and that his small differences were only
he variations between twro methods in both of which the arsenic
ientoxide and only that was actually determined.

Pentavalent Aisenic in Glass. — The above method of separat-
ng the two forms of arsenic was now applied to the glasses in
1 /. Ind. Eng. Chem., 4, 16 (191 2).

'52

JOURNAL OF Till-:

Tabls 4.
Showing the separation of given quantities of trivalent from pentavalenl

arsenic and the detcrrninat ion of tlie latter.1

As2Q3 taken

As206 taken

Mg2As207 found

Equivalent to As205

Variation from As20
taken

g-
o .0100

0.0023
0.0032
0.0024

O.OOOI

O. IOOO

0.0058
0.0075
0.0055

0.0003

g-

O.OIOO

0.01 16
0.0152
0.01 13

0.0003

0.0500
0.0581

0.0777

0.0575

0.0006

8-

0 .0100

o. 1 161
0.1565
0.1158

0.0003

t.
None

0 <x>,S*

o .0069

0.0051

0.0007

K

None
0 003J
f) ,005a
0.0038

o .0003

Table 3 and the pentavalent arsenic was determined in each
by the Gooch and Browning method. The results are recorded
in Table 5 . Duplicate determinations show a maximum variation
of 0.3 mg. and a mean variation of 0.1 mg. just as in the determina-
tions of total arsenic. In Table 6 are tabulated the amount of
arsenic pentoxide in each glass, together with the amount of
arsenic trioxide found by difference between the pentavalent
and the total arsenic. Evidently the major part of the arsenic
is in the form of pentoxide, indeed the trioxide is so small, ranging
from zero to a maximum of 0.09 per cent of the glass, that one is
inclined to suspect some systematic and tolerably constant
error between the two methods. It seemed quite improbable
that errors of this degree of constancy could be caused by spatter-
ing on the hot plate during the solution of the glass. Still some
losses in the beginning of the work, when fewer precautions were
taken, were traced to this source. Before going further, there-
fore, it was proved beyond question that the results were affected
by no such errors.

Effect of the Presence of Ferrous Iron in the Separation of
the Two Forms of Arsenic. — It occurred to us that the small
differences found between the total and pentavalent arsenic in
glasses might be due to the reducing action of ferrous iron. All
glasses contain a small amount of iron some of which may be
ferrous. Now if this should reduce a little of the pentavaler
1 Determined as Mg2As207.

AMERICAN CERAMIC SOCIETY.

753

Table 5-
tavalent arsenic in glasses determined iodometrically
As205 in milligrams per gram of glass.1

in duplicate .

Kind of
glass.

j Jena
/O3832

j Jena
IO340

I Jena
!0i44
I Jena
'O167

I Jena
\O211

l Jena
'O2122

\ Jena
/ O2071

G. iodine

solution

used.

4.72

5 06

6.23
6.23

4.82
4.19

5.38

6 .02

5.78
5 -92

10.18
10. 19

12 .16
12.23

Mg.
AS2O5
found.

2 .1
2-3

2.8
2.8

2 .1

1 -9

2.4

2.7

2.6
2.6

4-5
4-5

5 4

5-5

No. of

glass.

13

15 •■

Kind of
glass.

\ Jena
} O1209

S Jena
{O722

\ Jena
(O578

Jena
<?599

(P. P. G. Co.
1 Spectacle
[ Crown

(P. P. G. Co.
/ Green Plate

(P. P. G. Co.
j Dense Flint

G. iodine

solution

used.

IO.92
10.74

12.34
12.55

II .62
II .34

5.8l
5 93

1994

1978

3 69
3-51

18.37
18.25

Mg.
As20s
found.

4-9

4.8

5-4
56

5-1
50

2.6
2.6

9-9

9 •"

9.
9.0

:nic to the trivalent state during the solution of the glass by
)huric and hydrofluoric acids, this trivalent arsenic would
the same time be volatilized and therefore the pentavalent
:nic would always be found lower than the total arsenic even
:n the arsenic was entirely in the pentavalent state. The
)wing experiments were tried to test this point.
Inown amounts of ferrous sulphate were introduced into the
:ible along with measured quantities of arsenic acid and the
poration with hydrofluoric and sulphuric acids made as
)re.2

1 Determinations 1-9 inclusive were made with iodine sol. I; 1 g.
equiv. to 0.000446 g. AS2O5. Determinations 11-15 inclusive were made

1 iodine sol. II; 1 g. sol. equiv. to 0.000442 g. AS2O5. Determinations
•.2 inclusive were made with iodine sol. Ill; 1 g. sol. equiv. to 0.000494
s206.

2 Here, as in some previous results, the work was done before the more
rate iodometric method was adopted.

754

l< HJRNAL OF THE

.a o

« be
C cj

s.s

0&-

a a g

_

'1

to

»o

'

_

0

*■

c-

_o «

o

C

0

■*

«■>

«o

(N

-

o

"-J

00

1^

lO

M

0

-

o

■<*■

CO

00

*o

VO

r~

w>

-

o

<N

c

00

0

^C

■*

>o

VO

lO

M

o

'1

0

"0

-1-

o

CO

0

*o

•+

N

c

(S

O

<o

O

00

VO

VO

o

cs

a n

3 cS

o p

.S '5

B -g

O

c c

co a

O

■•=.

o £

>

|_!

w>

1— 1

_J

>

o

°!7i

C/J

=

o

—

a

>

-

-./

CI

>

a;

-

P

U

AMERICAN CERAMIC SOCIETY. 755

Taken 0.0097 g. FeS04.7H20 equiv. to 0.0025 g- FeO

Taken o .01 16 g. AS2O5

Found o .05 1 2 g. Mg2As20- equiv. to o .0113 g. As205

Error o . 0003

Taken 0.0193 g. FeS04.7H2C) equiv. to 0.0050 g. FeO

Taken o .01 1 6 g. As20.-,

Found 0.0153 g. Mg2As207 equiv. to. . 0.0113 g. AsO,

Error o . 0003

re we took as large a quantity of arsenic as we have found
any glass and much more oxide of iron than would be found
any but very special kinds of glass. It is evident that a quan-
y equal to 0.5 per cent FeO reckoned on a basis of 1 g. glass has
influence on the determination.

Effect of Ferric Iron on the Separation of Trivalent from
ntavalent Arsenic. — Further thinking on the real nature of
1 small differences which we call arsenic trioxide led to the
jgestion that the true values of arsenic trioxide in the glasses
ght actually be greater than the values found, because a part
it may have been oxidized to arsenic acid by ferric iron in the
Dcess of solution of the glass. The following experiment was
ide to cover this point.

3.050 g. Fe2(SO<i)3 was put into the crucible with 5 cc. 1 : 1
SO4 and heated a half hour or more on the hot plate to make
■e that any nitric acid, which the ferric sulphate sometimes
titains, was removed. The crucible was then cooled and 0.0503
AS2O3 was added. The crucible was covered and heated for an
ur or more. Then it was cooled again, 5 cc. HF was added
d the whole evaporated till white fumes appeared. The
;idue was tested rigorously with potassium iodide and hydrogen
[phide. No arsenic was found. Therefore all the arsenic was
latilized and none could have been oxidized by the ferric sul-
ate. Evidently then the presence of ferric iron does not in-
here with the separation of the two forms of arsenic by hydro-
oric and sulphuric acids.

Effect of Chlorine in Glass on the Accuracy of the Separation
the Two Forms of Arsenic— Most, if not all glasses, contain

756 JOURNAL OF THE

chlorine, doubtless in the form of chloride. In the Jena gla
analyzed by us, the largest quantity of chlorine found was 0.06
per cent. In an American optical glass 0.16 per cent was found.
Glasses made from material less carefully selected may contain
still more. According to Dr. E. C. Sullivan, chief chemist of
the Corning Glass Works, the largest quantity of chlorine which has
been found in any glass in that laboratory is 0.14 per cent. Now
since arsenic is volatilized when either oxide is heated with hydro-
chloric acid under favorable conditions, e. g., during the evapora-
tion in the silica determination in a glass, it is quite conceivable
that arsenic acid should be reduced by hydrochloric acid and
volatilized in the solution of a glass by sulphuric and hydro-
fluoric acids. (See footnote p. 742.) To test this point, weighed
quantities of sodium chloride were heated with given amounts
of arsenic acid, sulphuric and hydrofluoric acids. The crucibles
contained :

Glass free from arsenic 1 .00 g.

Arsenic acid solution, 10 cc, equiv. to 0.0058 g. AS2O5

NaCl=o .004 g. equiv. to 0.0025 g. CI

This was evaporated to dryness in a covered platinum dish.
5 cc. HF was now added and when the salts had dissolved, 5 cc.
1 : 1 H2SO4 was also added and the whole evaporated to white
fumes. The arsenic was then determined as usual.

11.64 %■ iodine solution used in titration; 1 g. solution is equiv-
alent to 0.000494 g. As206-

Taken 0.00580 g. AS2O5

Found o .00575 g- As205

Error o .00005 S- AS2O5

Again the following quantities of the same substances were used
in the same way.

Glass free from arsenic. 1 .00 g.

As2Os o .0029 g.

NaCl =0.0080 g. equiv. to 0.005 g-

H2SO« = 5 cc.
HF =5 cc.

5.97 g. iodine solution used in titration.

AMERICAN CERAMIC SOCIETY. 757

Taken o .00290 g. AS2O5

Found 0.00295 S- AS2O5

Error o .00005 S- Aso06

These experiments show conclusively that quantities of chlorine
rf the magnitude found in glasses do not interfere with the separa-
tion of the two forms of arsenic.

Influence of Platinum Dissolved from the Crucible on the
Separation of the Two Forms of Arsenic. — In the fusion required
n the determination of total arsenic a small amount of platinum
s dissolved which eventually is precipitated with the arsenic
sulphide. This platinum sulphide is somewhat soluble in am-
nonia and ammonium carbonate. Since platinum is an element
)f two valencies it is conceivable that reduction by hydriodic
icid and subsequent titration with iodine solution might consume
1 small amount of the latter. Two blanks were therefore made,
n which 3 g. Na2C03 and 0.1 g. KN03 were fused in a platinum
crucible with 0.7 g. and 0.5 g. pure quartz, respectively. Each
usion cake was then treated exactly as in an arsenic determina-
ion. A minute black precipitate brought down by hydrogen
iulphide was digested with ammonia and ammonium carbonate,
>xidized by hydrogen dioxide, reduced and titrated.

No. 1 required 0.16 g. iodine solution, equivalent to 0.08 mg. AS2O6.
No. 2 required 0.10 g. iodine solution, equivalent to 0.05 mg. AS2O5.

still another blank was made by heating sulphuric acid and hydro-
luoric acid in a platinum dish just as in the determination of
>entavalent arsenic. The amount of iodine required in titration
vas 0.18 g., equivalent to 0.09 mg. AS2O5. It is therefore obvious
hat the difference in the results obtained on any glass by the two
nethods cannot be accounted for by platinum derived
rom the crucible in fusion, nor can the difference be accounted
or by any other influence which may be eliminated by making
)lanks, for the blanks are of the same order of magnitude.

Separation and Direct Determination of Trivalent Arsenic. —
horn a purely analytical point of view, it is perhaps a question
>f no great interest whether the small differences which we found
)etween our series of determinations of total arsenic on the one

7.S*

JOURNAL OF THE

hand and pcntax ak-nt arsenic on the other really represented
the quantity of arsenious oxide present; but in its relation to the
chemistry of glass making, this question becomes a matter of more
importance.

We were unable to find any considerable error in our results,
but as we were still unconvinced of their meaning, we soughl
positive evidence in the matter by attempting to treat the glasses
with hydrofluoric and sulphuric acids in such a way that the volatile
material eould be caught and examined. In principle, of course,
the matter is perfectly simple, but the necessity of using hydro-
fluoric acid makes a glass apparatus unsuitable since the glass
might contain arsenic and the quantity of arsenic in question
is very small. Fortunately, we were able to get from the labora-
tory of the U. S. Geological Survey a small platinum still which
could be satisfactorily adapted to the purpose. The head of the
still, which is fitted to it by a taper joint, was made tight with

a rubber washer and brass clamps (Fig. i). The distillate
was caught in a caustic soda solution held in a receiver con-

AMERICAN CERAMIC SOCIETY. 759

nected to the still by an improvised gold tube and rubber stoppers.
The volatile products are swept out of the still into the receiver
by a current of air which enters the still through a small platinum
tube. The joint between this tube and the head of the still was
originally made with a perforated rubber stopper. This worked
well when the still was very gently heated, but if the distillation
was hurried the results were high, i. <?., some pentavalent arsenic
was carried over into the receiver. The losses were partly me-
chanical undoubtedly, and in part due, as we suspected, to the
reducing action of the rubber stopper which was at times badly
attacked. The latter was, at any rate, eliminated by replacing
the rubber stopper by a cork which was protected from the vapors
of the still by a thin, close-fitting gold ferrule which made a suf-
ficiently tight joint with the still. The mechanical losses were
successfully avoided by dropping into the still, with the reagents
and arsenical materials, a few pieces of crumpled platinum foil
and 0.5-1.0 g. of lead sulphate. The still was heated by the aid
of a radiator.1

For the most careful tests a gold bottle was substituted for the
glass receiver which was always found to be attacked and which
might have yielded a little arsenic to the solution.

With these safeguards the direct determination of arsenious
arsenic in known mixtures with arsenic acid gave very accurate
results. The determination is here a little more complicated
than in previous cases because the arsenite solution now contains an
excess of fluoride. If this solution were acidified and precipitated
hot with hydrogen sulphide, the hydrofluoric acid would attack
the glass container and might increase the quantity of arsenic.
The alkaline solution was therefore evaporated to near dryness
with a little hydrogen peroxide. This of course oxidized the
arsenic and permitted the separation of excess hydrofluoric acid
by fuming with sulphuric acid. (See p. 751.) The salt cake so
obtained was cooled, dissolved in hot water and filtered, after
which the arsenic was precipitated with hydrogen sulphide and
determined as usual. The results of the separation of small
quantities of arsenious acid from arsenic acid are given in Table 7.

1 See Hillebrand, "Analysis of silicate and carbonate rocks." United
States Geol Survey, Bull. 422, 31.

-(H)

JOURNAL OF THE

Tablh 7-

Separation by distillation of AsiOi from known mixtures with As.o., and
determination of the former.

Grams iodine- solution
required

Equivalent to As203 in
mg

As203 taken in mg

As206 taken in mg

Error.

0. 16

0.74

1 .70

2 .32

3.87

0.07

0.3

0.7

1 .0

1 .6

None

0.3

0.5

1 .0

15

1 16 .0

116 .0

58.0

11 .6

58.0

0.07

0.0

0.2

0.0

0. 1

452

1 -9

2 .0
580

o. 1

The necessity of refining the method to this point will be
obvious if the reader bears in mind its purpose, which was to
decide whether certain small differences between two methods
for determining arsenic really and accurately represented arsenious
oxide. The method was now applied to a chosen number of
the glasses. The results, which are tabulated in Table 8, show

Table 8.
Direct determination of trivalent arsenic in glasses compared with the dif-
ference between pentavalent and total arsenic.
AS2O3 in milligrams per gram of glass.

Jena Optical.

Pittsburgh

Spectacle

Crown.

1.

2.

7.

8.

9.

1 1.

13.

20.

Grams iodine solution

I -47

O.84

1 -37

O.60

O.91

T 28

0 28

2 .47

Equivalent to mg. As203

(A)

0.6

03

0.6

03

O.4

0.5

O.I

I .O

Difference between pen-

tavalent and total ar-

senic mg. As203 (B). . . .

0.9

O .0

0.7

0.4

0.3

O.7

O.I

O.9

Difference between A

0.3

0.3

0 . 1

O. I

O . I

O .2

0 .0

O. I

■

not only that direct determinations of arsenious oxide in these
glasses give positive results, but. that these results are of the same
order of magnitude as the differences in question. In five out of
1 1 g. iodine solution is equivalent to 0.000425 g. AS2O3.

AMERICAN CERAMIC SOCIETY.. 761

ight glasses tested the variation between the direct and indirect
etermination of arsenious oxide is only o. 1 mg. and in the other
hree the variation is only 0.3 mg. at the maximum.

It is obvious that the accuracy of the determination might
e considerably increased by taking larger portions for analysis,
ut as our supply of material was limited and the method had been
rought to a degree of precision quite sufficient for our purposes,
he investigation was pursued no further.

II. Determination of Boric Acid.
The direct determination of boric acid in silicates has hitherto
een a tedious process and for this reason this constituent is
requently not reported, or if known to be present is determined
y difference. Such a procedure is of course very unsatisfactory
ince all the errors made in the determination of the other con-
tinents fall upon the boric acid. In order to avoid this dif-
culty, we directed our efforts toward finding a satisfactory
nethod for the direct determination of boric acid in the Jena
ptical glasses submitted to us for analysis.

Wheiry's Method. — The method originated by Wherry1 and
nodified by Sullivan and Taylor2 seemed rapid and accurate
nd probably best adapted to our needs. We made repeated
ttempts to determine by the modified method the boric acid
n a glass containing about 0.5 per cent B203, making double
>recipitations with CaC03. The results were erratic, and while
ye have not done enough work on the method to say that some
nodification of our procedure might not materially improve the
esults, the necessity of washing out the boric acid in an alkaline
olution from bulky gelatinous precipitates is a fundamental
>bjection to it. The tenacity with which such precipitates
etain other substances is well known; then, too, we could not
ivoid the suspicion that with some combinations of elements
l substance with the characteristics of boric acid might form
nsoluble compounds which could not be removed at all. We
ire aware that the method is used in the laboratories of some of
he largest glass works and it is presumably valuable as a control

1 Wherry and Chapin, /. Am. Chem. Soc, 30, 1687 (1908).

2 Sullivan and Taylor, /. Ind. F.ng. Chem., 6, 897 (1914!.

JOURNAL OF THE

method, bul as a method of high accuracy it cannol be recom-
mended tor general use. Wherry himself luis expressed the same
opinion.

Electrometric Method. —Boric acid can be determined rapidly
and accurately by the electrometric1 method. In order to use
this method, in the analysis of glasses, however, the boric acid
must first be quantitatively separated from the other constituents
in the glass, especially from the arsenic. If the arsenic is presenl
as arsenic acid we have an acid whose strength2 naturally vitiates
the results, and, if present as arsentous acid, we have an acid
which behaves much like boric acid in the presence of the mannite'
which must be added to the solution when the electrometric
method is used.

Chapin's Method. — After the fruitless efforts which we have
recounted, Chapin's4 method was considered. We may anticipate
here by saying that for the determination of boric acid in silicates
of such varying composition as the glasses, the method is, certainly
at the present time, superior to all others. The description of
the method may give one the impression that it is very tedious,
but after the apparatus is once set up and the standard solutions
are made, the determination can be carried out as rapidly as most
of the analytical methods used in the analysis of silicates.

The method of Chapin is a combination distillation and titra-
tion method and consists in brief of the following procedure:5
The powdered silicate is fused with Na2C03 and the cooled melt
taken up with i : i HC1. The boric acid thus set free is volatil-
ized with pure methyl alcohol as methyl borate in the presence of
granular anhydrous CaCl2 which serves as a dehydrating agent.
The distillates are treated with NaOH, either N/2 or N/5, de-

1 J. H. Hildebrand, /. Am. Chem. Soc, 35, 861 (1913).

2 Abegg's "Handbuch," Vol. Ill, part 3, p. 539, and Vol. Ill, part 1, p.

29-

3 The large amount of alkali used up when AS2O3 in the presence of
mannite is treated with Ba(OH)2, using phenolphthalein as indicator, proves
this.

4 J. Am. Chem. Soc, 30, 1691 (1908).

5 The reader is referred to Chapin's description of the method for the
necessary details, J. Am. Chem. Soc, 30, 1687 (1908).

AMERICAN CERAMIC SOCIETY. 763

►ending on the amount of boric acid present. The NaOH com-
>ines with the boric acid and permits the removal of the methyl
lcohol by distillation. The alkaline borate is treated with HC1
nd heated under the proper conditions to expel C02. The solution
5 now ready for the final titration, which is carried out as follows:
fhe excess of HC1 is neutralized with NaOH, using para-nitro-
•henol to indicate the end point. Mannite is now introduced
nd the titration of the boric acid carried out using a few drops
f phenolphthalein as indicator.

Chapin tried out the method in the presence of such sub-
tances as fluorides, and the oxides of arsenic and zinc. These
ubstances are often present as essential constituents of glasses,
n every case his results show close agreement between the amount
f boric acid taken and found. We have subjected this method
o a rather thorough investigation and can in the main confirm
is conclusions.

Determination of the Blank. — It was considered advisable
0 learn if boric acid could be satisfactorily determined under
he usual laboratory conditions where one is forced to use glass
asks, such as Jena or Pyrex, which contain boric acid; and also
; our chemicals contained boric acid. In the three experiments
abulated below all the flasks used were made of Jena glass. Three
rams of Na2C03 were treated with 12 cc. of 1 : 1 HC1 in the
ecomposition flask and one gram of granular anhydrous CaCl*
>er cc. of liquid, added to the solution. The vapors of pure
nethyl alcohol were passed into the flask until about 200 cc. of
he distillate had been obtained. For the final titration we
sed Ba(OH)2T instead of NaOH, since the former can be prepared
nd maintained free from CO2 much more readily than the latter,
lie Ba(OH)2 was about N/10 and was standardized both against
IC1 and pure fused B2O3, using para-nitrophenol and phenol-
ihthalein as indicators, the titration for the second end point
ieing carried out as usual in the presence of mannite. On ac-
ount of the volatility of B2O3 in a hot solution the titration must be
arried out in the cold ; therefore if it is necessary to add water to

1 H. Will, Arch. Pharm., 225, 1101 (1887); E. Zschimmer, Chem. Ztg.,
5, 442-67 (1901); Sullivan and Taylor, ./. Ind. Eng. Chem., 6, 897 (1914).

7<>4

JOURNAL OF THE

the solution lor rinsing purposes such water must be free from Cm.
Undue shaking of the solution must also be avoided on account
of the sensitiveness of phenolphthalein to the COs thus intro-
duced. Under these conditions, both end points can be sharply
determined. The results of the three experiments mentioned
above are shown below:

Expt.

Ba(OH)i used
Cc.

Kquiv. to BjOa.
G.

i

2

3

o .20
0.25
0.20

O.OOO7

O OOO9
O 0007

From these results it is evident that a blank must be made under
the same conditions as those which prevail in an actual determina-
tion of boric acid. This blank may be due to several causes any
one or all of which may be operative: (1) Boric acid may be
present in the chemicals, or (2) it may be derived from the glass,
(3) finally, the difficulty of getting sharp end points in the solution
obtained after the distillation with methyl alcohol may also in-
troduce a plus error. We have noticed repeatedly that the second
end point in the final titration is never as sharp as in the titration
in the pure aqueous solution used for the standardization of the
Ba(OH)2. This point will be brought up again later on.

The aqueous solution obtained after removing the methyl
alcohol and acidifying were tested for B2O3 by the method of
Bertrand and Agulhon.1 We have found this method very
satisfactory for small quantities. Boric acid was shown to be
present in all our blanks.

Subsequently, we endeavored to determine the source of the
B203. The glass decomposition flask was replaced by a platinum
still, the glass condenser tube with one of silica glass, and the
evaporation of the distillate carried out in a platinum vessel.
All the other conditions were the same as those in the experi-
ments just described. The titration with Ba(OH)2,2 which was

1 Bertrand and Agulhon, Compt. Rend., 157, 1433 (1912).

2 The Ba(OH)2 used in these experiments was l/« the strength of the
Ba(OH)2 mentioned in the experiments described above.

AMERICAN CERAMIC SOCIETY.

765

rried as usual in the presence of mannite, showed that under
ese new conditions our blank amounted to 0.7 mg., which is
actically identical with results obtained above. We feel con-
lent therefore that no appreciable amount of boric acid was de-
red from the glass parts of our apparatus.1 It is probable,
wever, that this would not have been the case had fluoride
en present.

It was also found that a small amount of Ba(OH)2 was used
1 in passing from the first end point (para-nitrophenol) to the
:ond (phenolphthalein), when the titration was carried out in
stilled water free from C02. The amount of Ba(OH)2 used was
uivalent to 0.3 mg. It is evident therefore that the major
rtion of the blank is due to the boric acid contained in the
emicals. The amount of the blank obtained in the experiments
is constant enough to encourage us to continue our work and
termine the correction to be applied to our boric acid values.
In the following experiments known quantities of the boric
id used in the titration of the standard Ba(OH)>, together with
jrams of Na2C03 and the other necessary reagents, were used,
le boric acid was, as usual, volatilized with pure methyl alcohol.

Table 9.

tpt.

B2O3 taken. Ba(OH)2 used. B2O3 found.
G. Cc. G.

O.OI24
O.0434
O.0764
o. 1085

0.1363

3.85

12 .70
22.25
31 56
36.7O

O.OI34

o . 0440

O.O772

o 1094

0.1274

Excess.
G.

O.OOIO
O.OO06
O.OOO8

o 0009

O.OOII

the determination has been properly carried out the B203
ind is always greater than that taken. The results also show
it the excess is independent of the amount of boric acid taken.
e therefore felt justified in applying this excess (average = 0.9
j.) as a correction to the amount of boric acid found.

1 Our experience on this point does not confirm Mellor's statement.
: J. W. Mellor, "A Treatise on Quantitative Inorganic Analysis" (Griffin
Co., London), p. 586, footnote 2.

JOURNAL OF THE

in two experiments we found that the- amount of boric acl
determined was less thai] the amount taken. This was sliowi
to be due to the fact that the volume of the solution in the decom
position flask was larger than usual and therefore required mon
CaCls. ft is well known that the ordinary anhydrous CaCl
has an alkaline reaction. The additional amount of CaCl2 re
quired in the experiments just mentioned rendered the solutioi
alkaline and prevented the volatilization of part of the borii
acid. In carrying out a determination it is well, therefore, t<
determine the amount of HC1 required to neutralize the alkalinity
of a unit weight of the CaCl2 used and add sufficient HC1 to th
decomposition flask to insure an excess of acid.

This excess, however, should not be unduly large, else too grea
an amount of NaOH will be required for the neutralization o
the acid in the distillate. It has been shown above that ou
blank was caused almost entirely by B203 in the reagents, hence th
necessity of avoiding an undue excess of any reagent used in th
method.

Influence of Impurities in Alcohol. — Chapin has called at
tention to the fact that commercial wood alcohol will not answe
for the determination of boric acid. We have confirmed thi
and find that the difficulty is caused by a yellow-colored in
purity in the distillates which renders an exact determinatio
of the first end point in which para-nitrophenol is used, out ||
the question. Then, too, the second end point is so indistin<
that one can never be sure when it has been reached. In th
connection it was thought that possibly another pair of indicato
might overcome this difficulty. We did not succeed in findir
such a pair, but we did find that when pure methyl alcohol w;
used, methyl red and thymol blue1 yielded results which we
quite as good as those obtained with para-nitrophenol and phenc
phthalein.

One of the more common impurities found in methyl alcoh
is acetone. In order to find out if this substance was the cau
of the difficulty the following experiments were carried out :
1 Lubs and Clark, J. Wash. Acad. Set., 5, 614 (1915).

AMERICAN CERAMIC SOCIETY.

767

xpt.

B2O3 taken.
G.

Pure acetone in
methyl alcohol.

B2O3 found.
G.

Corrected.
G.

O.0358
O.0528

i per cent
5 per cent

O.0364
O.0538

OO355
O.0529

2

t (2), the second end point was indistinct and somewhat difficult
determine, which was not the case in (1). The experiments
ov/ that while it is not advisable to use an alcohol which con-
ins over 1 per cent of acetone, this impurity is not the cause of
e difficulty.1 Squibb's methyl alcohol (analytical reagent)
id Mallinckrodt's pure methyl alcohol gave excellent results.

Influence of Arsenic as As203. — It is well known that glass
ually contains arsenic. In the determination of boric acid in
glass relatively high in arsenic, part of the latter distills over
th the methyl borate; it was therefore deemed advisable to
vestigate the effect of this arsenic on the boric acid determina-
)ii. Varying amounts of pure As203,2 together with boric acid,
ire introduced into the decomposition flask along with the neces-
ry reagents. The arsenic carried over into the distillates was
itermined3 after titrating the B203.

Table 10.

xpt.

B2O3 taken.
G.

AS2O3 taken.
G.

B2O3 found

(corrected).

G.

AS2O3 found.
G.

O.0715
O.0702
00743
O.0795

O.OI 10
0 .0105
0 .0500
0 . IOOO

O.0715

O.0701
O .0767

Trace

2

3

4

0 . 00 1 1
0.0245
0.0568 (a)
0.0110 (b)

l (1), the amount of HC1 added was such as to neutralize all
ises and leave an excess of about 0.5 cc. cone. HC1. The amount
boric acid found agrees well with that taken, thus showing that

1 Mellor, loc. cit., p. 586.

2 vSee p. 746 for the method by which the As203 was prepared.

' See p. 745 for description of the methods used in making the analyses
arsenic.

768

|< OJRNAL OF THE

the accuracy of the determination has not been influenced by thi
amounl ofAsaOj taken. In (2), the amount of IIC1 in c
was about 1 cc. of cone. HC1. The accuracy has not been apj
preciably afTectcd, though the distillates contain an appreciable
amount of AsoO,). In (3), the amount of arsenic was increa ed
and the excess of HC1 was 2 cc. The end point in this case
was very indistinct and the amount of boric acid found was
greatly in excess of the amount taken. In (4), 3 cc. of cone.
HC1 were present in excess of the requirements. The boric acid
could not be determined on account of the constant fading out
of the second end point. The arsenic was determined separately
in the first and second distillates. The amounts are shown in
the table and are marked a and b.

It is evident from these results that if the amount of As20j
present is small and the excess of HC1 is only about 1 cc. the
determination of boric acid is just as accurate as if no As20j
had been present. The amount of AS2O3 used in the first two
experiments represents the maximum amount ordinarily found
in glasses. If, however, the substance to be analyzed contains
a large amount of AS2O3 the determination of B20.j becomes
inaccurate or impossible unless the AS2O3 is first oxidized to As20f
in the manner suggested below. The behavior of AS2O3 under these
conditions is not surprising when we stop to consider the ease with
which As2Oa is volatilized in the presence of HC1.

Influence of Arsenic as As205. — It also seemed worth while
to know whether pentavalent arsenic interferes with the accuracy
of the boric acid determination, especially as most of the arsenic
in glass is in that form. The As205 used in these experiment:
was prepared in the manner described on p. 746.

Expt.

B2O3 taken.
G.

AS2O5 taken.
G.

B2O3 found

(corrected).

G.

Arsenic as
AS2O3 in the
distillate.
G.

1

O.0762
OO745

O.0500
O . 0600

O.0765
O .0746

O .0018
0.0019

2

In both experiments 2 cc. of cone. HC1 were present in excess
The results show that the presence of relatively large quantitif

AMERICAN CERAMIC SOCIETY.

769

As2Os do not influence the accuracy of the boric acid determina-
n. It is rather interesting to note that some arsenic was
ried over into the distillate. This could not have been de-
ed from AS2O3 which might have been present in the As205,
?ause the manner in which the As205 was prepared precluded the
isibility of the presence of AS2O3. It is well established that in
: presence of concentrated HC1 a partial reduction of the
.O5 to AS2O3 takes place.1 The presence of the trivalent arsenic
hus accounted for. On the basis of these results it appears ad-
able, where boric acid must be determined in a substance
ich contains a large quantity of trivalent arsenic, to oxidize
latter in a solution made distinctly alkaline with NaOH before
tilling off the boric acid.

Since certain types of glasses contain fluorine, an experiment
5 carried out in order to learn if this element introduced any
turbing features. A solution containing NaF and B20,-j was
de alkaline with 20 cc. of N/2 NaOH, evaporated to dry-
is, acidified with HC1 and distilled as usual.

B»Oj taken.

NaF.

Ba(OH),. used.
Cc.

B2O3 found
(corrected).

O . I5OO

0.200

40.70

O.1517

e determination of the second end point was not as sharp as
lal, and the difference between the amount found and taken
greater than the experimental error. Chapin carried out a
lilar experiment, using, however, a smaller amount of NaF,
1 obtained very close agreement between the amount of B20;5
en and found. This indicates that the method is very accurate
en relatively small amounts of fluorides are present. The
thod is, however, useful even under the adverse conditions
)ur experiment.

Vhen small quantities of B2O3, say 2 mg. or less, are to be
ermined, it will be best to use the method described by Bertrand
I Agulhon,2 provided all the chemicals are known to be

1 Ahegg's "Handbuch," Vol. Ill, part 3, p. 539.

-' Bertrand and Agulhon, Compt. Rend., 157, 1433 (1913).

77<>

JOURNAL OF THE

free from BsO;).1 The method is a very sensitive one and mu
be carried out with great care.

Finally, we will present a few results showing the close agre
inent between the determinations when carried out by this meth<
and when the proper correction is applied, on different sampl
of the same glass. In the table below are tabulated a seri
of three experiments each on a glass relatively high in B20.-t ai
on one low in B2O3.

Table i i .

Glass.

Amount of

sample.

G.

Ba(OH)j used.
Cc.

fttOi found.
G.

Corrected.
G.

%of B*

I

2

| 0.5000
\ 0.3000
[ O.500O
[ I .0000
I I .OOOO
1 I OOOO

34-95

2I.IO

35 00

2 . 10

2 .20

2 .25

O.1311
O.0792
O.I313
0.0073
O.OO76
O.OO78

O. 13OI
O.0782
O.1313
0.0064
O.OO67
O OO69

26.0:

26 o<

26 .o<

0 6<

0.6

0 1

In the course of our work we have carried out about 50 defo
minations of boric acid by the Chapin method and have be
greatly impressed by its neatness and accuracy.

III. Determination of Some Other Constituents.

A considerable number of the constituents of glass may
satisfactorily determined in exactly the same way as is done
rocks. This is true of silica, chlorine and manganese, also
calcium, magnesium and the alkalies, except when boric acid
present.2 In the presence of boric acid only one modificat
of the usual procedure is necessary. (See p. 779.) But gl
contains other elements in addition to those considered in
foregoing pages, the separation and determination of which
rarely called for in rock analysis; or elements which, on accoi
of their quantity or their association with other elements, pres I

1 No doubt for very small quantities of B2C3, it will be well to avoid
use of glass apparatus or at least the use of such kinds of glass apparatui
contain B2O3.

2 Some glasses contain phosphoric acid; these, of course, demand sp(f
methods. We have had no experience with them.

AMERICAN CERAMIC SOCIETY. 771

ecial difficulties to the analyst. We shall make no attempt
cover the whole field — only such portions of it as have come with-
our own experience, where we have found difficulties and have
ined some knowledge of the accuracy of certain methods.

1. Determination of Iron in Optical Glasses. — Since a very
tie iron colors glass appreciably and lowers its transmission,
ly a very small amount is permissible in an optical glass. The
:urate determination of this iron demands several precautions;
re must be taken to exclude iron from all outside sources and
avoid loss of iron by occlusion in those precipitations which
:essity demand before iron can be determined. First a special
nple of about 5 g. of the glass should be prepared by crushing
agate instead of steel; a crucible of the purest platinum should
used in the various operations requiring one, especially when
! fusions are made. Finally a suitable blank must be made.
Two or three grams of the sample are decomposed with about
:c. 1 : 1 sulphuric acid and 5 cc. hydrofluoric acid just as in the
termination of pentavalent arsenic. Cool, transfer the residue
a beaker with water and boil till any insoluble sulphate becomes
erable. Filter into a 150 cc. flask, wash, and save the pre-
itate which may contain as much as 50 per cent of the total iron.
2cipitate the filtrate hot with hydrogen sulphide. Thus arsenic
d the remaining lead1 are removed. If zinc is present, proceed
:ording to Waring,2 viz., add to the solution before precipita-
n with hydrogen sulphide a drop or two of methyl orange,
itralize with sodium carbonate and add 0.3 to 0.5 cc. excess
50 per cent formic acid. In a solution of this acidity zinc
phide is precipitated in a denser form than in an
:aline solution. Waring draws up the conducting tube above
: level of the liquid and corks the flask tight, thus leaving the
uid under a slight pressure of the gas over night. The filtrate
m the zinc sulphide which ordinarily contains the major
rt of the iron is boiled to expel hydrogen sulphide and pre-
itated by boiling with pure ammonia and a few cc. of hy-
)gen peroxide.

1 The amount of lead not precipitated by sulphuric acid is appreciable
;n the sulphate is filtered immediately after precipitation.

2 J. Am. Chcm. Soc, 26, 4 (1904).

77-' JOURNAL OF THE

Ii the k';iss contains lead the insoluble sulphates should first
be leached on the filter with a hot concentrated ammoniacal
solution of ammonium acetate which should previously be filtered
if it contains any iron whatever. When the lead sulphate has
all been washed through by this means, the residue, whet her
it contains barium sulphate or not, should be incinerated in thi
platinum crucible and then fused for a short time with a weighed
portion of pure potassium or sodium bisulphate. From the
cooled mass the iron is extracted by hot dilute sulphuric acid and
precipitated like the major portion. Both the very small pre-
cipitates of ferric hydroxide are then filtered and washed on
the same filter, incinerated in the pure platinum crucible and fused
with a weighed portion of pure potassium bisulphate for a few
minutes at a comparatively low temperature (under a red heat).
All the iron is easily dissolved. The remainder of the determina-
tion is done according to Hillebrand,1 i. e., the fused cake is dis-
solved in water (no sulphuric acid is needed here) and the solution
reduced by hydrogen sulphide; the small precipitate of platinum
sulphide filtered off", and the hydrogen sulphide expelled from
the filtrate by boiling in a current of carbon dioxide which is
ivashed by a solution of copper sulphate. This operation is very
conveniently carried out in a glass-stoppered wash bottle.

The cold solution is titrated by a very dilute solution of potassium
permanganate with the aid of a weight burette. The permanga-
nate is best standardized by sodium oxalate. The blank should
be carried out as nearly as may be like the determination itself:
especially should the same quantities of hydrofluoric acid and potas
sium bisulphate be used and the fusions should be conducted in
a similar manner as to time and temperature.

Some results are given in Table 12 which may be of interest
in showing the minute quantities of iron in optical glasses and the
remarkable uniformity of the Jena glasses.

2. The Determination of Zinc. — Zinc may of course be de
termined in the filtrate from silica, or preferably in the same
sample as arsenic and lead.- One or two grams of the glass an

1 United States Geol. Survey, Bull. 422, 107.

2 When barium is present the procedure on p. 777 may be followed.

AMERICAN CERAMIC SOCIETY.

773

3* x S

a 2 5.

P o '

-l — I— (

u i— l c«

n> _. rs

S a. o

r* 2. <-■

O rt

_ <-h 3

O 3

s -

p

3'

a

4*

u

P

8

o

3

>

Ov

j
n

0

ID

>

o

ID

n'
p

3
re

r*

p

n

K

-

P

p

J2 C

o

o

IT)

o

£3

o

c

o

~1

-

00

p

w

ID

^

JOC

o o

S«

^ *1 ^

1 1 .•*

p 93 n>

•-t pi a:

<"3 3 3

g 1 8

3 3 <gi

2 H *

71 H o

a, *-> -^

g g -

a = 5-

=T P 3

§ O O

ooS

*■ *■ 03

o <n n

l/l In

9 Ofi

• e 3.

2. a. p

C C 7?

p c. <

~ £■ n

w O 93

o' o' -

^ 3 O

p p

o 3 £

a' a

"»«! O

pT c

: 11 o

p sr

7? ft

p p

C-

tfl

(« .

O 0 0

0 0 fo

O " ^

'« 4* <W

—

-h to Ln

O M

H

o o o

0 0 N)

b m i

M 6. <w

N)

- o o

Os o\ '

00

O O 0

0 0 K>

0 M -t>

- o^"

w

U 10 -1-

o> o '

o

<— i

0 0 o

O O to

b '« 'm

M tM «

4^

O

M k) Ul

Os vO

•o

M

o o o

O O Oo

0 M ^J

Ih >© <w

Cn

9

k> C\ ~j

Co 0 '

p"

0

r»

o o o

O O U

0-4-

- as <ra

OS

H U Ol

CN to

Oo

0 o o

O O Oo

0 m ^i

m i W

c

►0 0*0

0

0 O 0

O O Oo

O M "J

- 00 <R

—

« Cn +>

Oj ~-l

p

O

0 o -

0 M gj

0 +> Oo

M +. W

-

>

+> 4* Oo

0 O

3

0

T

n

o o «

0 - to

0 t" O

H H IK

IsJ

5

4- N4 M

O OC

o

■o

r.'

o o o

0 O to

bii^

h M*i

;-

O

OJ Oi -

-J 00

K3

IT

0 o -

0 « Oj

b oo -

- iv. K

+v

M 00 4-*

Osl ^4

vC

3

.-

m

«

o

M

3

(0

774 J< »' RNAL OJF THE

weighed out, decomposed by hydrofluoric and sulphuric add i

the product evaporated to white fumes. The residue is cooh
transferred to a beaker and diluted. If all is soluble the so
tion is used directly for the zinc- determination. If barium
present the solution should first be boiled for some time to reiu
the sulphate filterable, while if either barium or lead is prese
the solution should be set aside for the night to insure compl<
precipitation. The precipitate is then filtered on a Gooch crucil
if lead is present, otherwise on paper; the washing in either a
following the well known methods. The filtrate and washinj
are taken for the determination of zinc.

In this solution the arsenic may be precipitated first by 1
drogen sulphide or both arsenic and zinc may be precipitat
together by the same reagent. In either case before zinc
precipitated, the acidity of the solution should be reduced accoi
ing to Waring,2 and the zinc sulphide brought down in a 1
and slightly acid solution. If the sulphides of zinc and arsei
are precipitated together, the precipitate is washed with 1
water, the funnel covered with a watch glass and the zinc s
phide dissolved by pouring cautiously under the cover a lit
hot i : i hydrochloric acid, while the filtrate is caught in a beak
After the filter has been thoroughly washed with hot wat
it is well to dissolve the arsenic sulphide by pouring a few
colorless ammonium sulphide directly on the filter and washi
into a separate beaker. A very little zinc sulphide may
left and it is safe, therefore, to treat the filter again with a lit
hot, dilute, hydrochloric acid, adding the filtrate to the zinc so
tion. After the hydrogen sulphide has been boiled out of 1
latter the zinc may be satisfactorily determined by seve
methods.

Sullivan and Taylor3 add a little more than the required quant
of sulphuric acid, evaporate to dryness, heat to redness and we:
the sulphate. Their results were very good. In some of <
determinations we have made a double precipitation of the z
by sodium carbonate in very slightly alkaline solution, accord

1 Excepting of course the final washing with alcohol when lead is pres

2 Loc. cit.

3 J. Ind. Eng. Chem., i, 476, (1909).

AMERICAN CERAMIC SOCIETY. 775

readwell,1 filtering on a Gooch filter and heating inside a
porcelain crucible2 to the highest temperature of the Bunsen

er.3

000 g. pure metallic zinc by this method gave 0.1245 g.
equivalent to 0.1007 g. zinc. The following duplicate

ts on glasses were obtained by separating the zinc in the

ler described, precipitating the redissolved sulphide as

mate and weighing as oxide:

ZnO found.

ena optical glass No. 3 .

ena optical glass No. 6.

enp. optical glass No. 8.

ena optical glass No. 15.

' 1 .00 per cent
. o . 96 per cent
' 8 .46 per cent
.8.61 per cent
' 1 .14 per cent
. 1 .24 per cent
1 .42 per cent
. 1 .46 per cent

little more accurate but not quite so rapid as the carbonate
od is the electrolytic method. If the latter is to be employed,
:hloride solution obtained by dissolving the zinc sulphide is
umed with sulphuric acid in slight excess, cooled and diluted
>o cc. Pure caustic soda is now added till the precipitated
lydroxide is redissolved; then acetic acid in quantity sufficient
■ecipitate and again dissolve the hydroxide. The solution
idulated with 0.5 cc. glacial acetic acid, diluted to 150 cc.
electrolyzed at room temperature. To secure good results
following points must be heeded: In the first place zinc
at be precipitated directly on platinum without alloying with
latinum somewhat and roughening its surface. Also the cath-
nust be smooth if a uniform, firmly adhering deposit is to be
ned.4 To obviate these difficulties the platinum cathode is
ully coated with copper. This may be done by electrolyzing

Treadwell, trans, by Hall, II, 4th ed., Wiley & vSons, X. Y., p. 142.
Ibid., p. 27.

An electric oven of suitable type was not at that time available.
H. J. S. Sand, /. Chem. Soc. London, Trans., 91, 383 (1907).

::<>

JOURNAL OF THE

a copper sulphate solution containing 2 per cent nitri<
;i suitable current density. We used a rotating gauze eatri
of about 85 cm", available surface, gradually increasing the cur!
from 0.25 to 1. 00 amp. over a period of 20 minutes. The cathi
was washed with water, then successively by alcohol and it
and dried in a current of air at room temperature. I'ina
in the electrolytic precipitation of zinc, a current under 4 m
must be used; otherwise hydrogen will be liberated, the 08
will be spongy and will not cling to the electrode. Heeding
above precautions, using the rotating1 gauze cathode de < til
and electrolyzing with a current which was gradually rai
from 0.25 to 2.5 amp. at a maximum of 4 volts, thirty minu
proved sufficient for the deposition of over 0.300 g. of zinc, \\h
adhered well to the cathode. It is unwise to precipitate mi
more than this amount on a cathode of this area (85 cm2.), sii
larger deposits show a tendency to flake off. The zinc is was!
without breaking the current and dried exactly like the cop;
deposit mentioned above.

A standard sample of zinc from the U. S. Bureau of Standa
gave the following results:

Zinc taken.
G.

Found.
G.

I

O.3227
O.3227
O.2154
O.2154

O.323O
O.3232
0 2 I 59

2

3

4

O.2154

In the first three experiments the zinc was simply dissoh
in such an amount of sulphuric acid as to leave about 1 cc. pi
H2SO4 in excess, after which the procedure was exactly as p
scribed above. In the fourth experiment the zinc was precipita
as sulphide, dissolved and further treated like the other thi
A comparison of the carbonate and electrolytic methods on
American optical glass follows:

1 See Gooch and Medway, Am. J. Sci., [4] 15, 320 (1903); Ibid.,
56 (1904); Exner, J. Am. Chem. Soc, 25, 896 (1903).

AMERICAN CERAMIC SOCIETY. 777

Carbonate method

Electrolytic method.

nC) found.

1 1 7 per cent
I .12 per cent

1 . 19 per cent
1 . 20 per cent
1 .25 per cent

hough we have not tried it, the electrometric method for zinc
Biehowsky1 appears rapid and accurate and would have
ded advantages where many determinations were to be made.

Determination of Lead and Barium Occurring Together. —

1 glass which contains both lead and barium we have found
est to determine the two together as sulphates; then in a
rate portion to determine the barium directly and thus
iin the lead by difference. The lead may also be determined
rely but this is more difficult. The details follow:
weighed portion of glass is decomposed by sulphuric and hy-
luoric acids and evaporated till all the hydrofluoric acid is
en out. The pasty mass after cooling is transferred to a beaker
i jet of water, diluted to about 150 cc, boiled and left to stand
■ night. The filtering, washing and heating of the precipitate
carried out exactly as with pure lead sulphate,
arium and lead sulphates cannot be separated by ammonium
ate solution ; some of the former is sure to dissolve with the
sulphate. Barium is therefore determined in a separate
ion of glass which is fused with sodium carbonate. The silica
>parated as usual, arsenic and lead are precipitated as sul-
les by Waring's method,2 the alumina, iron, etc., are pre-
tated by ammonium sulphide, and in the final filtrate
1 which the hydrogen sulphide has been boiled out the barium
etermined as usual. In accurate work it should be noted
: the barium sulphate, obtained either by precipitation from
te solution, or by decomposition of the glass with sulphuric and
rofluoric acids, always occludes some soluble sulphate and the
fht is therefore too high. In two barium glasses which con-
ed less than half a per cent of Na20, the barium sulphate

1 J. hid. Eng. Client., 9, 668 (1017).
; hoc. cit.

778 JOURNAL OF THE

obtained in the latter way contained the following impurita
0.6293 g. BaSO< «l;iss No. 7 contained 0.0052 k- NajSOi.

0.7140 K- BaSO< glass NTo. 8 contained 0.0061 «. Na.SO,.

The errors in the percentage of BaO in the above determinatic
would amount to 0.17 and 0.20 per cent, respectively.

In case a direct determination of lead is desired, the gl;
should be fused with sodium and potassium carbonates at
lowest possible temperature to avoid volatilization of lead oxi
The loss is then very small, if appreciable. The silica is n
removed in the usual way except that the operator should ws
it more carefully with hot water to remove the lead chlori
In any event it is wise to look for lead in the silica when the lat
is volatilized with sulphuric and hydrofluoric acids. The resic
should be carefully fused with a very little soda and added
the lead chloride solution. The latter is now made sufficiently a<
and precipitated hot with hydrogen sulphide. After standii
the lead sulphide is filtered on a small filter, washed thorough
and transferred by a jet of water to a small casserole in whi
it is dissolved in nitric acid and finally evaporated to wh
fumes with sulphuric acid. The further treatment of the p
cipitate is carried out as usual, and the little lead sulphide whi
clings to the filter on which it was washed must be burned at a I
temperature in a small tared porcelain capsule and finally fun
with a drop or two of sulphuric acid. The weight of the le
sulphate thus obtained is added to that of the main portion. S
livan and Taylor2 obtain results on lead 0.2 to 0.3 per cent lo^
when it is determined in the filtrate from silica than when it
determined by decomposing the glass with oxalic and hydrofluc
acids. Our results are similar when sulphuric is substitui
for oxalic acid. Thus a glass which gave by the lat

method > PbO, when fused with soda and heal

35.26 per cent J

further as described above, gave 35.13 per cent PbO.

A second glass, which by the first method gave 41.64 per 0]

PbO, gave 41.31 per cent by the second. The lower resr

1 For the method used in this determination see Allen and Johns'
/. Am. Chem. Soc, 32, 588 (1910).

2 Loc. cit.

AMERICAN CERAMIC SOCIETY. 779

iine4 by the second method are not necessarily due to the
itilization of lead in the fusion ; they are quite as likely due to
hanical losses in the more complicated procedure which this
hod demands.

Separation of Alumina from Barium and Calcium. —
ical glasses not infrequently contain alumina in amounts
jing from 0.5 per cent to 5.0 per cent and sometimes much
e The precipitated alumina accompanied as it is by almost
ron is very difficult to filter,1 yet when an alkaline earth is
ent it should be twice precipitated. One of us prefers to pre-
:ate the first time cold with ammonium sulphide, by making
filtrate from zinc and arsenic, which already contains hy-
[en sulphide, barely ammoniacal and allowing it to stand over
t while the other makes a double precipitation with ammonia,
rever precipitated, every filtrate from the alumina should
ivaporated to a small volume on the water bath to make

no alumina is lost. If any appears it is filtered on a sep-
e filter.

is possibly not so generally known as it should be that am-
ium hydroxide which has stood in glass for any considerable
: invariably gives too high results when used to precipitate

and alumina. Its freedom from silica and alumina or any
r impurity which might interfere with the determination
lid be insured, after which the ammonia should be kept in
sine or gold bottles.

Analysis of Glasses Containing Boric Acid. — Chapters on
analysis of glass including borosilicates, though not found
■dinary text-books on analysis, are given by Classen3 and by

1 Since iron is determined in a separate sample it would doubtless be
r to add to the solution, before the alumina is precipitated, a small,
ured volume of a dilute standard ferric chloride solution since, as is
known, an alumina precipitate containing several per cent of iron gives
ouble of this kind.

! The ammonia water used in this laboratory is made as follows. Lique-
immonia gas is purchased in large steel cylinders from which small dis-
ting cylinders of similar kind are filled. The gas is washed in a little
r and absorbed in distilled water contained in a gold bottle packed in ice.
1 "Ausgewahlte Methoden der Analytischen Chemie," Braunschweig,
, I, 609.

78o JOURNAL OF THE

Lunge.1 The determination of boric acid itself we have already

dealt with (p. 761). For the removal of boric acid where it would
interfere with the determination of other elements, the same
principle is utilized, i. e., volatilization of the boric acid by methyl
alcohol partially saturated with hydrochloric acid gas.2 To
guard against the precipitation of boric acid along with alumina,
lime or magnesia, the boric acid should be removed in the prog-
ress of the silica determination. When the solution has been
evaporated to dryness the second time to remove the last of the
silica, cool the dish and add 25 cc. or so of the methyl alcohol
solution, cover the dish and evaporate again on the steam bath.
Repeat the operation, finally rinsing off the cover into the dish.
Lunge recommends evaporating 3 or 4 times with methyl alcohol,
but so far as our experience goes, two evaporations are sufficient.
Classen and Lunge both determine the alkalies in glass by
Berzelius' method, in which the boric acid is volatilized as fluoride
along with silica. More satisfactory is the method of J. Lawrence
Smith.3 In the decomposition of the glass by this method the
boric acid is retained, and must be removed in a subsequent
operation. When in the course of this method the chlorides
of ammonium and the alkalis have been evaporated to dryness
on the steam bath, 10 cc. of the methyl alcohol solution should
be added and evaporated with the dish covered. One repetition
is all that is necessary, in most cases, for the complete removal
of the boric acid. In case of doubt it is a simple matter to re-
peat the operation on the weighed chlorides.

6. Determination of Moisture. — In the complete analysis of
a glass it must be remembered that fine mineral powders gener-
ally absorb moisture from the atmosphere, the quantity absorbed
by any particular mineral depending upon the fineness of the
powder, i. e., upon the extent of its surface. It may be surmised
that glasses on account of their alkali content would be specially
liable to absorb water. On this account the glass should be

1 "Chemisch-technische Untersuehungsmethoden," Berlin, 1904, I, 666.

2 Jannasch and Heidenreich, Z. anorg. Ckem., 12, 208.

3 Am. J. Set., [3] I, 269 (1871). The method is described also by
Hillebrand, loc. cit., p. 171; H. S. Washington, "The Analysis of Silicate
Rocks," 3d Ed., Wiley & Sons, N. Y., p. 193 and by Treadwell-Hall, p. 496.

AMERICAN CERAMIC SOCIETY.

781

crushed for a moment only, either in a hardened steel or an agate
mortar, and then screened, the operation being repeated till
sufficient powder is obtained. In this way very fine powder is
almost entirely eliminated and only a minimum of moisture is
absorbed. Optical glass powders prepared in this way gave the
following losses when heated for a few minutes in a covered crucible
over a Bunsen burner :

Moisture in Glass Samples.

Jena Glasses.

American Glasses.

No. of glass

1

2

3

4

5

6

12

13

14

15

I

2

3

4

Per cent loss

O.06

O.09

0.20

O.08

O.03

O.14

O.15

O.08

O.08

O.16

O.06

O. IO

O.06

Two other glass samples, which were inadvertently left in a me-
chanical grinder for several hours and were consequently very
fine, lost on heating 0.68 per cent and 0.88 per cent, respectively.
The following experiments made for another purpose prove that
the loss is unquestionably water. A certain glass, which lost
0.10 per cent when heated over a burner, was fused in a Gooch
tubulated platinum crucible1 with dry sodium carbonate, the
moisture expelled by dry air and absorbed by calcium chloride.

From 1 g. glass was thus obtained 0.9 mg. HaO = 0.09 per cent.
From 2 g. glass was thus obtained 2.2 mg. H20 = 0.11 per cent.

7. Gases in Glass. — Low results in the analytical summation
of certain glasses led us to consider the question whether gases
occur in glass in appreciable quantity. Judging from its com-
position and the materials used in the manufacture of glass we
supposed carbon dioxide was most likely to be found. We selected
for the test a glass which from its high content in lime and alkali
we believed would have a stronger tendency to retain this gas.
14.73 Per cent CaO, 10.37 Per cent Na20 and 0.44 per cent K>()
were found by analysis. For the determination of carbon dioxide
we used Borgstrom's modification2 of the ordinary method, in
which he decomposes the silicate by a dilute solution of hydro-

1 Am. Chem. Jour., 2, 247 (1880); Hillebrand, Bull., 422, p. 78.
- Z. anal. Chem., 53, 685 (191 4).

782 JOURNAL OF THE

chloric and hydrofluoric adds. The gaseous mixture expelled from
the decomposition vessel was driven through the usual train of
absorbents. Borgstrdm made three blanks in which his soda
lime tube gained from 0.8 to 1.6 mg. This amount he finds
is no less when hydrochloric acid alone is used in the decomposi-
tion of the mineral.

Since the material which we had to test was glass, glass was
excluded from certain parts of our apparatus. We used a plati-
num decomposition vessel and a water-jacketed lead tube as a
condenser. Using Borgstrom's acid mixture in a blank experiment
with this apparatus we found a gain of 1.4 mg. in the soda-lime
tube. When 3 g. glass was heated in the same way for the same
time the gain was 1.7 mg., showing that the carbon dioxide in
this glass was small if appreciable. More than that, it is impossible
to say on account of the large blank which the method involves.

Far more satisfactory is the following experiment carried out
by our colleague Mr. E. S. Shepherd who has had unusual ex-
perience in work of this character. Mr. Shepherd very kindly
tested the glass in question for gases. A large sample of the glass
powder was weighed into a platinum boat which was heated in
an evacuated silica glass tube at 12000. The gases were pumped
out and analyzed. Before the gases were measured they were
drawn through an absorption tube containing silver oxide1 and
potassium carbonate. 0.0025 g. S02 equivalent to 0.87 cc. under
standard conditions was removed. This is not included in the
analysis.

6.577 g. glass gave 6.55 cc. gas measured under standard condi-
tions. The analysis gave the following volume percentages:

02 = 64 . 2 per cent

CO2 = 24.2 per cent

CO = 3.5 per cent

H2 = 3.9 per cent

N2 = 4.1 per cent

99 9
A small fraction of the C02, less than 1 cc. at the most, may
come from the potassium carbonate through decomposition by
1 According to Mr. Shepherd, this method of separating S02 originated
by him is not perfectly sharp but better than any other.

AMERICAN CERAMIC SOCIETY. 783

SO2. Uncorrected the total carbon dioxide amounts to 0.05
weight per cent of the glass, while the total gas amounts to 0.15
weight per cent. If the reader cares to speculate on the source
of the different gases it may be stated that the glass contained
0.74 per cent S03 which might give rise to the sulphur dioxide
and a very little of the oxygen. It contained no arsenic. Whether
nitrate was used in the manufacture we do not know. It was not
an optical glass. The obvious fact is that glasses may contain an
appreciable amount of gas most of which would probably not
be included in an ordinary analysis in any form.

Composition of the Jena Optical Glasses. — This paper is
chiefly concerned with analytical methods, but it will not be
out of place to recall the purpose for which the work was originally
undertaken, namely, to get information which might be of value
in establishing the optical glass industry in this country. The
Jena glasses were selected for investigation because of their
world-famed excellence; because they are especially well known
to scientists through the valuable publications of O. Schott
and others ; and because a large number of them were immediately
available to us.

Table 13 contains nine analyses of Jena optical glasses. Nearly
as many more were actually made but those given are typical.
About half of them are virtually complete, so far as we know,
except for the determination of gases which they probably con-
tain. The summations of the remainder would be raised several
tenths of a per cent by the addition of magnesia, sulphur tri-
oxide, chlorine and moisture.

One American optical glass, 3a, is included in the table. It is
evidently a copy of No. 5 and is practically as good as its proto-
type. This glass is also of interest because it contains no ar-
senic. The most obvious information to be gained from the
analyses of the Jena glasses is that they must have been made
from very pure raw materials, for the impurities lime, magnesia,
sulphur trioxide, chlorine and iron are all present in insignificant
quantities. There is practically no iron and extremely little
chlorine and sulphur trioxide which in certain glasses are liable
to cause milkiness. It is also a safe inference that the pots used
must have been not only very low in iron but very refractory,

|« 'i RNAL OF THE

O

TO

-,

■j

w

o

-J
a

;

<

H

o

M

c

-t

-

NO

o

K

4

1/-.

'

**i

00

•'

00

•O C

1 1

c c

*3

00

o ■ •

C

?

-r

C 0 •

• - CC

-'. y / /

c c

8

N

0

"■J vO

1/3 C

-r c

■ O o to c

8s

00

•O

c

1/3

vC '-/.

f! g< io N

N

a

N £ 1/1 -1-

-

C

"! °°,

in

-1-

0 0

• 0 vc

0

<*3

■*

-t

vC 0

'O ic, jj Q ej

M « c o d

- T|

/

o

o

C

— ■ oi

1/3

0
c*3

i*J

c

• — r-

- 1

0 0

00

r->

sC

C* ff

c*3 f!

'0 io r^

3 N

1 ^

on

M

Os

■+ c

• - O

0 f) 0) M 4J

■ - c

i «

9

CM

5

c

>c oo
— >d

0

-r

»o

c c

■ c c

o n 0 oc 5
+ 2 Z

"i 0

8

sO

sn

1/3 -t

- a

•+ o

T|

CC

_

__

IT

o

u~, C

M C

' Ol • M g

o

-t

oo

c

-1

;:

c c

• 0 0

-t- Z

1/3

0>

VC 00

0>

— 00

v— -—

CN

(N

00 «*•

o

(S f<*

so

T|

0
c*s

^C sc
— ' \6

SO

5 5
Z Z

■ -t *o

• CN u~

cd

^ C c S ° 0 0

6 5 • £ «i d o

Z o Z •+

0

0

00
0s

N

'O -+■ 0

r-i oo

r^ f

VO <N

c

1 r<

cm O «-"

• r-- r-

O aj •--

0 c

c

1 r0

"-•

3 «

VO

0 0 0

■ - vo

0 ■ o • "■

0 o

0

2

_ sO

-r

"I ' *

1 y

o

>o 00

0 00 <N

vO«0>or^-1-aj<u(u4>

-r -

oc

-+

c

M

X

CN O ""

C ~ 00

o

o c C C S

• 3 o 5 g

M Z Z ZH

o o

o

r~

5

_' -r

vC

0 0 O

O >C f

o

o o

o

lO u-

"* o

vO — O ai

<0 ■* a; *o o

^- M

o

r»

n

5

1^ _

r- -1- - 0

^t- C C 0

i/3

o to o a

0 * r-|

O vO r; Ov to 0 0

C - £ 0 Os c c

/-, (N

IN

0

Os

Os

Os

*

00

M

(N O X

>* -

c

•*

c-.

lO CV (N O □

O -+ fO 0

0 »/; «-" r- n«

o o

0

>o

c

to O
— ■*

sC

os os c o o

O 00 00

0

n ~ O O 0

0 " z z z

o o

0

OS
Os

c

S =i

u

2

.Q

3
C3

.

E

In

8

3

3

J3

3

0

E

°

3

^j

o

3
0

J3
0

a
C

C
c7

'c

c

1 <J

<

c

0

<

"S

—

c

c

e

r

0 «

c

c;

PC

1

9

S

c

-

-

c
<

C

c

AMERICAN CERAMIC SOCIETY. 785

:or the glasses are almost free from alumina except where it has
)bviously been added to the batch.

The authors wish to express their thanks for the loan of a number
>f platinum utensils by the chemical laboratory of the U. S.
geological Survey, which materially facilitated the work.

SUMMARY.
Arsenic.

1. An accurate method for the separation and determination
>f both trivalent and pentavalent arsenic in glasses is described.
1) The separation depends on the volatilization of the trivalent
arsenic as AsF3 when the glass is heated with hydrofluoric and
ulphuric acids, while the pentavalent arsenic remains in the
esidue. (2) The latter is determined by precipitation as sul-
fide which is then oxidized to arsenic acid, reduced by hydriodic
icid and titrated with a standard iodine solution. (3) The
rivalent arsenic is determined by difference between the pen-
avalent and total arsenic. The determination was controlled
>y direct determination with the aid of a platinum still (4) The
otal arsenic is determined by fusing the glass with sodium
ind niter, removing the silica and excess nitric acid by evaporation
rith sulphuric acid and subsequent filtration; and finally pro-
ceeding as outlined in (2).

2. These methods for arsenic in glasses are generally applicable
o substances in which the arsenic can be transformed into sul-
fide without loss, and are highly accurate.

3. A comparison of the iodometric method and the magnesium
>yroarsenate method for arsenic in glasses is made. The former
las the advantage in accuracy, and also in speed except where
>ccasional determinations are called for.

Boric Acid.

1. For the determination of boric acid we have found that
^hapin's method is very reliable and yields highly accurate re-
ults.

2. It has been shown that in order to obtain very accurate
esults, a "blank" must be made and the value applied as a cor-
ection to the amount of boric acid found. The correction is small
ind for ordinary work can be neglected.

786 AMERICAN CERAMIC SOCIETY.

3. The accuracy of the method is very appreciably affected
by relatively large amounts oi" arscnious acid but not by arsenic
arid. Boric acid can therefore be satisfactorily determined in
the presence of large amounts of arsenious acid by oxidizing the
solution with H2O2 after making it distinctly alkaline with NaOH.

4. Relatively large amounts of fluorides appreciably affect
the accuracy of the determination but do not seriously impair
its usefulness for ordinary work.

Other Determinations. — Our experience with the following
cases in glass analysis is detailed: (1) the determination of
the minute quantities of iron in optical glass; (2) the separation
and determination of zinc; (3) the separation and determination of
lead and barium occurring together ; (4) the separation of calcium
or barium from relatively large quantities of aluminum occurring
with almost no iron; (5) the determination of those elements
in boric acid glasses with which the boric acid interferes.

Attention is called to the universal presence of hygroscopic
moisture in powdered glass samples.

Some data by Mr. E- S. Shepherd on gases in glass are given.

Geophysical Laboratory,

Carnegie Institution of Washington,

Washington, D. C, January 15, 1919.

HE CONDITION OF ARSENIC IN GLASS AND ITS ROLE
IN GLASS-MAKING.

By E. T. Allen and E. G Zies.

The condition of arsenic in glass has been a mooted question
/ith glass-makers. O. Schott1 of the Jena glass works regarded
11 the arsenic in optical glass as present in the form of arsenate
mich, he believes, is a product of the reaction between the ar-
enious oxide and the alkali nitrate of the batch. Where nitrate
5 not added to the batch it is Schott' s opinion that the arsenic
> entirely volatilized from the molten glass; in other words that
lass never contains arsenic in the form of the lower oxide. On
tie other hand, W. Rosenhain1 of the Chance Glass Works reasoned
rom the general chemical behavior of arsenic that glass must con-
ain only the trivalent form.

F. Gelstharp2 found experimentally that when arsenious oxide
pas fused with sodium carbonate in an open crucible at 940 °,
. part of it was volatilized, and, of the remainder, the major
tart was oxidized to arsenate. When arsenious oxide was heated
rith calcium carbonate to 10900, again a portion of the arsenic was
Dst, while the rest was all oxidized to arsenate. Thus it appears
hat the air alone is capable of oxidizing arsenious oxide when thus
leated with lime or soda. Gelstharp found that even sodium
ulphate oxidized arsenious oxide when fused with it in a covered
rucible, as was shown by the formation of sodium sulphide,
le further states that the Pittsburgh plate glass was found by him

0 retain 56 per cent of the arsenic put in, entirely in the form of
rsenate.

Our own results show that in all the glasses tested, both plate
,nd optical glasses, the major part of the arsenic present exists
n the pentavalent state but that nevertheless a portion exists

1 the trivalent state.3

1 Hovestadt, "Jena Glass, etc." Translation by J. D. and A. Everett,
tfacmillan and Co., New York, 1902, p. 412.

2 Trans. Am. Cerant. Soc, 15, 585 (1913).

3 See the preceding paper in This Journal.

7ss Jl >URNAL OF THE

Role of Arsenic in Glass-Making. Arsenious oxide is used
in glass-making with two distinct purposes. In some cases it is sup
posed to "whiten" the glass. With this function we are not con
cerned in this paper.

Arsenious oxide is also used in fining or removing the small
hubbies from glass; it causes the glass to "boil," in which opera
tion the small bubbles are apparently swept out by larger ones.
Other substances which give off gas at high temperatures are used
at times for the same purpose — wood, potatoes, etc. Just what
the chemical behavior of arsenic in the fining process may be is
little understood. vSchott1 suggests that it may depend on the dis-
sociation of arsenates at high temperature attended by the libera-
tion of oxygen. As Schott has elsewhere expressed the view that
glass cannot retain arsenic in the trivalent condition, this dis-
sociation should result in the liberation and loss of arsenious oxide
as well as oxygen.

Very accurate methods for the determination of both oxides
of arsenic now enable us, with the aid of information from the
manufacturers, to compare the condition and quantity of arsenic
in the finished glass with that of the batch. In the table is
given the history of the arsenic in three American optical glasses
made at the factory of the Spencer Lens Company at Hamburg,
N. Y. The mixing of the batches was very carefully superintended
by Dr. C. N. Fenner of the Geophysical Laboratory, from whom
wTe obtained the samples and the composition of the batches
from which the percentages of arsenic in the finished glass could be
calculated. The history of each pot is knowrn in detail. The table
also contains data on two plate glasses. From the batch of each,
both plate and optical glasses, we have calculated the percentage
of arsenic (as trioxide) which would be found in the finished
glass, on the assumption that no arsenic, boric oxide, or alkali
was volatilized in the furnace; in other words, that only moisture,
carbon dioxide, and oxygen (from KNO3) were lost. Losses
of boric oxide and alkali would of course raise these figures a trifle,
but probably not by an appreciable amount. The percentages
are given in the table. The table further contains the quantities

1 Doelter, "Handbuch der Mineralchemie," Leipzig, 1, 861 (1912).

AMERICAN CERAMIC SOCIETY.

789

of trivalent and pentavalent arsenic which we have found in each
of the glasses by methods detailed in the preceding article in This
Journal. From these figures the loss of arsenic was readily cal-
culated. All the glasses have certainly lost arsenic, though of
course it is impossible to say at what stage of the process it was
lost. The green plate was the only glass of the batch to which
niter was not added, and here the loss of arsenic is distinctly
larger. With this exception the losses are of a similar order of
magnitude. Concerning this loss of arsenic, an observation made,
by R. L,. Frink1 is of considerable interest. He found, in the
bubbles or "seeds" of a glass, crystals which were identified both
microscopically and chemically as arsenious oxide. So far as
it goes, this observation supports the hypothesis of the dissocia-
tion of arsenic pentoxide and the volatilization of arsenic tri-
oxide. Thus it appears that arsenic tri oxide is oxidized at low

Table.
Showing the condition of arsenic in the finished product and the loss of arsenic
in the making of five American glasses.

Found.

Taken.

AS2O3.

%.

Lost in

•%of

glass.

Lost in

%of
As:03.

Kind of glass.

AS2O5.
%■

AS2O3.

Total ar-
senic cal.
as AS2O3.
%.

"T" Medium

flint containing

42 per cent PbO

O.48

O.38

0.05

O.38

O. IO

2 I

"P" Light flint

I- i

containing 38

percent PbO.. . .

O 40

0.35

O.O5

0-35

0.05

12

"V" Baryta flint

containing 38

percent PbO and

6 per cent BaO

O .26

O.23

O.03

O.23

O.03

I I

n J

Spectacle crown

I .09

O.99

O.08

0.93

O 16

15

'

O.36

O.18

O .09

O 24

O.I2

33

1. Optical glasses manufactured by the Spencer Lens Co. at Hamburg,
N. Y.

2. Optical glasses manufactured by the Pittsburgh Plate Glass Co.

1 Trans. Am. Ceram. Soc, 17, 798 (191 5).

79o AMERICAN CERAMIC SOCIETY.

temperature and the product formed is stable enough to remain
until a high temperature is reached and the glass becomes fluid ,
when it slowly dissociates into oxygen and arsenic trioxide, both
of which aid in the fining.

Further consideration of this interesting question must he
postponed until we have more data on the behavior of arsenates
and arsenites at high temperatures.

The authors wish to express their thanks to the authorities
of the Pittsburgh Plate Glass Company and the Spencer Lens
Company, and also to Dr. Fenner for the information and material
which they very kindly placed at our disposal.

Oi:< iphysical Laboratory,
Carnegie Institute of Washington,
Washington, D. C, January 15, 1919.

THE PARTIAL PURIFICATION OF ZIRCONIUM OXIDE.

By A. J. Phillips, Emporium, Pa.

Zirconium oxide is a very desirable refractory for use at high
temperatures, its high melting point as well as its low coefficient
of expansion causing it to be very satisfactory with respect to
checking and spalling upon rapid heating and cooling; likewise
its low thermal conductivity renders it very desirable as an in-
sulating material in certain kinds of work.

In the course of some work on spark plug bodies a considerable
quantity of the oxide was required for experimental purposes,
the specification calling for less than o.i per cent of Fe203. A
search was therefore made for a method which would remove the
iron oxide alone, the material thus purified and still containing
silica, alumina and oxide of titanium being sufficiently pure for
the purpose.

The crude zirconia secured from the Foote Mineral Co. of
Philadelphia was received already ground to a fineness of 75
per cent through the 100-mesh sieve and recourse was made to
the effect of sintering it with various fluxing agents. Sintering
with sodium carbonate gave an incompletely decomposed product
which on leaching with weak acid filtered very slowly and still
contained too much iron oxide to be satisfactory; furthermore
the amount of manipulation involved was excessive.

Roasting with ammonium chloride removed very little of the
iron as chloride and roasting with sodium chloride to 15000 C
gave a product containing considerable amounts of fixed alkali
and an iron oxide content of 1.5 per cent. Repeated treatments
with sodium chloride did not lower the iron oxide content, while
the alkali content increased with each treatment.

A search of the literature showed a large number of methods
reported for the decomposition of zircon, a rather complete
bibliography being given by J. W. Mellor.1

In order to compare several of these methods of decomposition
1 "Quantitative Inorganic Analysis," Vol. 1, p. 499.

792 I' 'i RNAL l >l; THE

a sample, usually m grams, was incorporated with a "decomposer,"
molded into a small cylinder and fired in a direcl Same to 1200
[300 ° C. Alter tlif firing the ignited mate-rial was ground,
treated with a 10 per cent solution of hydrochloric acid for two
hours and filtered. A weighed portion of the residue was then
fused with sodium bisulfate and the iron after reduction with
hydrogen sulfide, titrated with permanganate. In the following
table is given the percentage of iron oxide remaining after several
different methods of treatment.

Table 1 .
Composition. Percent PesOi.

1 1 part ZrOs + i part Na,.S()4 + '/s part C. i .70

2 1 part Zr02 + 1 part Na2S04 + 1 part C. . . . 1 .79

3 1 part ZrO-2 + excess of Na2S o .88

4 1 part Zr02 + 1 part NaCl 1 .68

5 1 part ZrOa + 1 part NaCl + l/a part C 1 .45

6 1 part ZrO-2 + 1 part NaFl o .92

7 1 part Zr02 + 1 part NaFl + l/s part C o .79

8 1 part Zr02 + 3/i part NaFl + lA part Na2S04 0.91

The use of sodium sulfide as a flux was not as satisfactory as
the results appear, since further treatment did not seem to have
any effect on the remaining iron oxide. Further work was then
done with the fluorides as they had the best effect, although two
well-known methods of fluoride decomposition were not tried;
one using a mixture of sodium hydrate and sodium fluoride, the
other acid potassium fluoride, as both methods would involve
the use of containers on account of the fluxing of the mixture.

A consideration limiting the use of various reagents was found
in the amount of zirconia rendered soluble by leaching with hy-
drochloric acid in each case. In all of the treatments the leachings
were reduced with hydrogen sulfide and boiled with a solution
of sodium thiosulfate, in order to precipitate the hydroxides of
zirconium, titanium and aluminium. It was necessary to re-
duce first with hydrogen sulfide, since before the liberated sulfur
dioxide reduced the iron to the ferrous condition, the precipitate
formed absorbed or occluded ferric iron which was not subsequently
reduced and removed by the sulfur dioxide — even after the re-
duction the precipitated material was not entirely free from iron

AMKRICAN CERAMIC SOCIETY. 793

oxide. It was therefore thought important to weigh the residues
in each Case after the acid treatment and ignition, in order to
determine how much was rendered soluble and would require
further treatment.

Tabu; 2.

Per cent I'er cent
Composition. residue. FesOa.

1 1 part ZrOj + >A part KB Ft, 58 .29 0.70

2 1 part ZrO, + 'A part (>A NaFl + '/, Na2S207) . . 73 .53 o .85

3 1 part Zr02 + 72 part (3A NaFl + 'AC) 66 .62 o 79

4 1 part Zr02 + 1/i part (3A Na2S207 + 'A C) 81 .00 0.96

S 1 part Zr02 + 'A pt. (2.5 pts. CaFl2 + 1 pt. H3BO3) 64.40 1 .21

6 1 part Zr02 + 'A part (5 CaFl2 + 1 Na2B40;) 52.62 1.15

7 1 part Zr02 + 'A part (5 NaFl + 1 Na2B407) 63 .40 o .62

8 1 part ZrO, + 'A part ('A Na2S207 + VaNa^O/) 61 .20 0.69

Table 2 shows the composition of the material ignited; the per-
centage of material remaining after igniting, leaching with acid
and again igniting; and the percentage of iron oxide remaining
after the treatment.

The first mixture fused and it was necessary to run it over in
platinum; the others were brittle and pulverized readily All
of the mixtures containing sodium pyrosulfate filtered poorly —
due to the presence of much gelatinous material. In general the
sodium fluoride-borax mixture was the best, but even its behavior
was not entirely satisfactory as evidenced by the presence of 0.62
per cent residual iron oxide. In cases where calcined zirconia
is ground and dissolved in hydrofluoric acid in order to effect
subsequent precipitation as potassium zirconium fluoride, the
use of this mixture would be advantageous since no container
is required for the fusion, whereas, when an acid fluoride is used,
a graphite or carborundum crucible is necessary.

The possibilities of acid treatment alone were then considered,
since Wedekind1 has recommended continued digestion with this
acid and Rosenhain2 states that the impure material readily
stands a temperature of 16000 C while the refractory qualities
are much improved by digestion with hydrochloric acid to re-
move a part of the iron oxide.

1 Berichte, 43, 290-7 (1910).

2 Met. Chem. Rug., 15, 674 (1916).

794 JOURNAL OF THE

For the acid tests, 10 grains of powder, which passed the ioo-
mesfa sieve, were digested with 50 ce. of acid for three hours on a
strain plate at ioo° C and the iron oxide in the residue determined.
In addition the effect of a small amount of pressure or at least
tlu- prevention of the volatilization of the acid was effected by
placing the powder and acid in a small bottle with a spring cover
and heating it in boiling water. In two cases the acid was first
saturated with chlorine gas. The results were as follows:

Per cent FetOi
Compositions. in residue.

i 10 grams powder, 50 cc. HC1 1 .52

2 10 grams powder, 50 cc. HC1 in bottle 1 .40

3 10 grams powder, 50 cc. HC1 sat. with CI 1 .48

4 10 grams powder, 50 cc. HC1 in bottle sat. with CI.. . . 1 .34

A few other tests were made involving the use of HC1 ; in one of
them the zirconia was acid-treated after having been heated to
900 ° C for three hours in a current of hydrogen. The results
were not satisfactory. In another case dry hydrochloric acid
gas was used instead of hydrogen and while the iron oxide not
removed by digestion with acid was reduced to 0.62 per cent,
yet this was considered still too high.

Chlorine gas was then tried on a small scale by passing it over
the ore in an electrically heated tube. The results were so en-
couraging that a gas-heated furnace was built in which the ore
was heated in an upright cylindrical muffle, while chlorine was
introduced at the bottom through a piece of glazed pyrometer
tubing.

The simultaneous use of chlorine and carbon as explained by
Thomsen1 causes a great reduction of the temperature at which
the iron oxide is decomposed. In the succeeding work, therefore,
the crude zirconia was mixed with about 4 per cent of finely pow-
dered petroleum coke, molded into small balls and heated in the
muffle in an atmosphere of chlorine to 900 ° C.

The question of retorts in which to conduct the chlorination is
an important one and probably a dense, completely vitrified
material resembling chemical stoneware would be most satis-
factory. The cylinders first used were jiggered from a mixture
1 "Thermochemistry," p. 357.

AMERICAN CERAMIC SOCIETY. 795

of 55 per cent grog passing an 8-mesh but retained on a 30-mesh
sieve, 35 per cent Florida kaolin and 10 per cent Tennessee ball
clay. While withstanding the rapid heating and cooling employed,
they were entirely too porous, the chlorine penetrating and coming
through the walls as shown by their whitening. A better mixture
was later secured, it being a glass-pot mixture of Laclede Christy
bond clay, Illinois and Newark kaolins, Arkansas and Highlands
fireclays with 60 per cent of 12 -mesh grog composed of North
Carolina kaolin and calcined Highlands fireclay. This cylinder
became red with repeated use, showing the penetration of iron
oxide, and apparently grew denser with time. Porcelain jars
were tried but cracked with rapid temperature changes and por-
celain inlet tubes for the chlorine became very porous even though
not heated to any extent.

To further increase the density of the cylinders they were
painted with a mixture of sodium silicate and clay which on heating
fused to a smooth glaze. On the external surface the glaze coating
was slowly volatilized, but on the inner surface a dense, white
salt coating was formed which did not spall to any appreciable
extent.

It was soon found that zirconium chloride was being volatilized
out of the furnace along with the iron chloride, for on dissolving
some of the sublimate in dilute HC1, neutralizing, saturating
with SO2 and boiling the solution, a considerable precipitate of
hydra ted zirconia was secured.

Prescott and Johnson,1 state that zirconia with charcoal is
volatilized as the chloride in a current of chlorine.

In order to recover this material a cylindrical reverberatory
type of furnace was built with a long cooling chamber. It was
also thought that passing the products of combustion over the
material would maintain a reducing atmosphere which might
result in cutting down the amount of carbon required in the raw
mix. However, the fact that the chlorine sank to the bottom of
the cylinder and was swept out by the current of gases kept the
balls at the top from being properly chlorinated unless a large
excess was used; therefore a return was made to vertical chlorina-
tion, an up-draft furnace being used and the products of com-
1 "Qualitative Chemical Analysis," p. 203.

796

Jl >i k\.\l. OF THE

bustion passed through the mix. The construction of the furnace

is shown in Pig. I.

With this furnace the top was closed and the gas and air intro-
duced near the top, the draft being upward through holes in the
bottom of the cylinder. Owing to the limited space, combustion
was very imperfect and it was necessary to keep the tip of the

AMERICAN CERAMIC SOCIETY. 797

turner partly outside of the furnace. The result of this partial
ombustion was that a considerable quantity of water vapor was
armed and condensed at the joint A, so that the clay luting
round the joint was dripping wet during over half the run, and
rater vapor with HC1 gas issued from the stack throughout
he run. The effluent contained considerable iron oxide as shown
y its red color as it dried on the outside ; however, an examination
f the material showed that little or no zirconia was present
nd it was suggested that external heating be used and that the
hlorine be passed through water. This was finally the method
dopted, the chlorine being passed through a bubble tube near
he furnace. Cold water was not nearly so efficient as hot water
l the bubble tube in retarding the choking of the apparatus with
olatilized iron chloride and probably for larger installations the
ljection of low pressure steam along with the chlorine would be
lost satisfactory.

In chlorinating with external heating and exposure of the mate-
ial to the air at the top of the furnace, peculiar incrustations
ke "Pharaoh's serpents" collected over the top layer, whereas,
1 cases where up-draft and a stack were used, the sublimate was
eposited as a fine powder upon the walls — the color in most
ases being an indication of how much zirconia had been volatilized
long with the iron oxide.

An interesting fact was that apparently the size of the speci-
lens made no difference, as balls from 1/i in. to 1 in. diam. were
hlorinated at the same rate and the necessity for the fine grinding
sually specified in decomposing this material was entirely ob-
iated — the raw material with only 75 per cent passing the 100-
lesh sieve being successively used.

The penetration of the chlorine was shown by the fact that
1 every case where a ball was white on the outside, it would also
e white on the inside right to the core. In a few cases some of
le balls at the top of the cylinder were white on the inside and
ellow or coated with yellow crystals on the outside. This was
yidently caused by iron chloride being volatilized lower down
1 the hotter part of the furnace and being deposited on the
x>ler material near the top.

798 JOURNAL OF THE

The question then arose as to whether or not it would be possible
to chlorinate the material without first mixing with a binder
and molding into bulls. Very probably, however, this would
be possible only if a rotating drum were used since there is always
some iron chloride which collects at the top layer of material and
which would sooner or later choke the furnace and shut off the
How of chlorine.

It would be possible then to press the crude zirconia into bricks
with a little binder, pile them in checker work in a furnace and
chlorinate them in place — if it were not for the fact that chlorina-
tion removes the iron oxide bond in the material and after ehlorina-
tion it is very mealy, porous or punky. The oxide being heavy,
the bricks would then fall to pieces or at least crush those at
the bottom of the pile. Even by firing to cone 35, which was
done in some cases in calcining the finished material, a very
slight bond was developed, the specimens being porous and not
permitting much handling.

In order to secure more quantitative figures as to the loss of
zirconia, two samples were chlorinated in an electrical tube-
furnace under the same conditions except that in one case the
chlorine was passed through cold and in the other, boiling water.
The analyses as given below show (1) for cold and (2) for boiling
water.

1. 2.

Per cent. Per cent.

Si02 15 .52 II .49

AI2O3 2 .44 I .87

Fe203 o .04 o .03

CaO 0.68 0.49

P2O5 0.05 0.03

Na20 0.06 0.05

K2O 0.09 0.08

Zr02 80.91 85.93

Ti02 0.18 0.16

If we take the following analysis (3) of the raw material and
calculate out the ignition loss and the oxides removed by wet chlo-
rine, viz., Fe2C>3, Mn02, P2O5, and Ti02, we obtain a theoretical
composition as shown in (4) :

AMERICAN CERAMIC SOCIETY. 799

3. 4.

Per cent. Per cent.

Si02 10.62 1156

A1203 i .71 i .86

Fe»03 2 .90 ....

CaO 0.42 0.45

P206 0.65

Na20 o . 04 o . 04

K2O 0.06 0.06

Zr02 7925 86 . 3 1

Ti02 1 . 8 1

Ig Loss 2.78 ....

By comparing analysis 4 with 1, about 5 per cent, and with 2,
less than 0.5 per cent ZrC>2 was lost in chlorination.

In explaining the course of the reactions, we might suppose that
the chlorides of iron and zirconium would be formed and volatilized
it the same rate with a loss of zirconium chloride at least equal
to that of the iron chloride. However, Mellor1 states that if a
:ertain resistance is to be overcome before a reaction can be
started, the process which requires the least energy for its initia-
tion will be produced in preference; furthermore, if the two re-
actions have actually started side by side, the relative velocities
:>f the two reactions will determine which one takes place to the
greatest extent. Iron chloride volatilizes at ioo°C while the
zirconium chloride volatilizes at or about 400 ° C ; so it is probable
:hat there is a preferential formation and volatilization of the
ron chloride. As the temperature is gradually raised we have
the establishment of a reversible reaction, FeCl3 K > FeC^ + CI,
3r, according to Burgess,2 two independent components form the
system which is divided into three phases, FeCl3, FeCb, and CI.
rhe system is therefore mono-varient and admits a curve of dis-
sociation temperatures. Therefore at any temperature T,
equilibrium corresponds to a tension P of the chlorine atmosphere
which is determined by a knowledge of the temperature T. If
:he rate of chlorine introduction is increased so that the pressure
Decomes greater than P, the FeCl2 again absorbs CI until the
Dressure equals P and if the chlorine is reduced, FeCl2 is reduced

1 "Chemical Statics and Dynamics," p. 325-6.

2 "Thermodynamics and Chemistry," p. 155.

800 AMERICAN CI'] RAM It' SOCIETY.

until the pressure takes its former value. In other words, aij
equilibrium is constantly tending to In- formed.

Now with the introduction of water vapor the equilibrium is
disturbed; there is a partial decomposition of the vapor into hy
drogen and oxygen which reacting with the chloride tends to the
formation of a new equilibrium containing among some other
products, oxide of iron and hydrochloric acid gas. If we apply
the same considerations to the- zirconium chloride we shall es
tablish a system containing some zirconia and hydrochloric aci
gas. There would thus be no advantage in the introduction o
water vapor if it were not for the fact that iron oxide mixed with
carbon is volatilized as the chloride in a current of hydrochloric
acid gas while zirconia is not affected thus; so that as we hav
an excess of carbon present the conditions are favorable for th
removal of the iron oxide alone.

Thus as the action to a large extent depends on hydrochloric
acid gas, it might seem advantageous to use it throughout in
decomposing the material, but in the tests made the decomposing
action of the hydrochloric acid gas was not sufficient to remove
the last 0.5 per cent of iron oxide, the more energetic chlorine
being required to break up the combination in which the iron
oxide exists in the crude zirconia.

It is obvious that the treatment is applicable not only to the
purification of crude zirconia, but also to bauxite and flint for
the removal of iron oxide.

In conclusion we beg to acknowledge our indebtedness to Mr
A. V. Bleininger of the Bureau of Standards, for suggestions
and assistance in outlining the work.

STRENGTH TESTS OF PLAIN AND PROTECTIVE
SHEET GLASS.1

By T. L. Sorey, Washington, D. C.

This paper deals with strength tests of plain sheet glass and
ne type and make of protective glass. The protective glass,
mich will be designated as Type A, consists of sheet glass to one
ide of which is firmly attached a thin sheet of celluloid. The
itention is that the glass shall be used with the celluloid or treated
ide toward the person to be protected. This is one of several
inds of protective glass which are to be investigated and there-
ore at present no statement can be made as to the value of this
lass compared with other protective types. However, the
ffect of the thin coating of celluloid on the mechanical properties
f the glass is interesting, and the presence of this coating enables
is to study the manner in which sheet glass is fractured by various
Drees.

The samples were subjected to firing tests, impact tests and
ross bend tests as follows:

Firing Tests.

The sheets of glass used for this test were 8" by 10" and of three
hicknesses, l//, Vie", and l//. The 1/8" and 7/ thicknesses
fere plate glass while the 3/i6" thickness was tested in both a
•late and a special blown window glass. One lot of glass of each
hickness was plain and another was Type A protective glass.

The samples were clamped firmly to a board frame with bear-
rigs on rubber strips 3/16" thick. One set of bearing strips was
iear the edges of the glass and another around a rectangular
role, 2V2" by 5", at the center of the sheet. The samples were
hen perforated by bullets fired from a regulation army rifle at a
istance of fifty feet.

The following observations were made from the results of the
iring tests:

1 By permission of the Director, Bureau of Standards.

JOl'RNAI. ()!• THE

A Plain glass:

All samples were completely broken and shattered.
B. Type A protective glass:

i . The glass remained in plaee with the exception of a small
area, in the immediate vicinity of the point of entrance, which
was pulverized and carried through with the bullet.

2. The thinner the sheet the fewer the cracks.

3. A smaller hole and fewer cracks resulted when the bullet
entered perpendicularly to the surface than when it entered at an
angle of 45 °.

4. Fewer cracks resulted when the bullet entered from the
treated side.

vMA /v

k

7

\

Fig. 1.

AMERICAN CERAMIC SOCIETY.

803

5. The effect of the shock was less on blown window glass than
n plate glass of the same thickness.

6. When the cracking was the most extended as in a few of the
lickest samples, the plate bent at the middle but was held
)gether by the celluloid.

\ \ Y v

Fig.

7. In handling and cleaning the samples, the celluloid was
isily scratched or marred if there was any grit on the surface.

The effects of bullets fired perpendicular to the surfaces and
gainst the uncoated sides of the Type A samples are shown in
igs. 1 to 4. These are typical fractures.

So)

JOURNAL ()!• Till;

Impact Tests.

The samples used in this test were of the same- size and kinds
as those used in the firing tests, except that no 3/]6" plate glass
was tested. The specimens were placed in the frame used in the
firing tests and this was supported solidly on the steel frame of an
impact testing machine of the pendulum type. The pendulum

HM -

Fig. 3-

consisted of a il/4 pound flattened steel ball at the end of a 24
stiff wire arm. The hammer had a spherical striking surface
with a radius of i1//.

In making these tests the pendulum was first allowed to swing
through 50, then io°, and so on, the angle of swing being in
creased each time by 50 until fracture occurred. However, a

AMERICAN CERAMIC SOCIETY.

805

11 through 90 ° was not sufficient to break some of the specimens,
which cases the hammer was dropped repeatedly through 90 °
itil the piece was fractured. The angle or the number of
petitions at 90 ° was recorded On several of the samples of
otective glass the test was continued until there was complete
icture of both glass and celluloid.

7ypc A

Fig. 4.

The energy of impact in foot pounds of a swinging pendulum,
hen impact occurs as in this case perpendicularly below the point
suspension, is calculated by the use of the following formula:

E = 1 — cosine angle A X L X W,
in which

Angle A = angle with vertical made by the pendulum arm
at the time of release.

8o6

JOURNAL OF THE

I. length of arm in feet.

W = weight of hammer in pounds.
It is impossible to figure the effect of repetitions of the same
blow or blows of increasing severity — due to lack of data on the
mechanical properties of glass. If the time required to alter the
apparatus had been available, the energy of impact necessary to!
fracture the strongest samples could have been determined byl
using a heavier hammer or a longer pendulum arm.

Tablk i.

Results of impact tests.1

Nominal thickness
of glass.
Inches.

Description.

Average
angle at
fracture.
Degrees.

Average

energy at

fracture.

Ft. lbs.

1/8

Plain

425

O.66

1/8

Treated side exposed

50.0

O.89

1/8

Glass side exposed

45 0

0.73

3/i6

Plain

90. o2

2 .46

3/i6

Treated side exposed

65.0

1 44

3/i6

Glass side exposed

90. o3

2 .46

1/4

Plain

62.5

1 -35

1/4

Treated side exposed

70.0

1 .64

1/4

Glass side exposed

82.5

2.17

The chief results of these tests are shown in Table 1. The
following observations were made at the time of testing:

1. The plain glass was broken into many small pieces. Th<
treated samples cracked but the pieces were held together by th<
celluloid coating. Only a few small splinters were loosened im
mediately under the hammer.

2. After the first fracture occurred in the treated samples
only two or three more blows were required to break the celluloic
and splinter the glass.

3. The cracked area was smaller on the thinner samples.

4. Fewer but more extended cracks were caused when the blov
fell on the treated side than on the glass side.

1 Each of the values given in the table is the average of two tests.

2 One sample did not break with 170 repetitions at 90 °.

3 Repeated an average of 8 times. Energy of impact figured simpl;
as of 90°.

AMERICAN CERAMIC SOCIETY.

807

Attention may be called to the fact, as shown in Table i, that
the 3/i6" uncoated blown window glass was stronger than the 1//
plate glass. Also, that the amount of energy required to cause
the first crack in the glass was not materially increased by the
celluloid coating.

Cross Bend Tests.

The samples used were 2" by 12" and of three thicknesses,
as in the preceding tests. The Vs" and 1/i" thicknesses were
plate glass and the 3/i6" thickness was blown window glass.

In making the tests, two wooden bearing blocks of triangular
section were placed on the platform of a small counter scales so
as to give a span of 10 inches. A strip of glass was laid on these
and the load applied to the middle of the sample by means of a
rod screwed downward through a steel yoke placed above the
scales The lower end of the rod was provided with a bearing
block of semi-circular cross-section resting in a ball and socket
joint. Thin bearing strips of rubber were interposed between the
glass and the three blocks.

Table 2.
Results of cross bend tests.1

Nominal thickness.
Inches.

Description.

Actual

average

thickness.

Inches.

Average
load at
failure.

Pounds.

1/8

Plain

O.142

20.5

1/8

Treated side up

O.I45

17.5

1/8

Treated side down

O.I52

31 -5

3/l6

Plain

O.I75

36 .0

3/l6

Treated side up

O.186

25.0

3/l6

Treated side down

0.190

46.5

1/4

Plain

0.229

44-5.

1/4

Treated side up

O.241

38.5.

1/4

Treated side down

O.242

60. o-

The results of these tests are shown in Table 2. Fracture
Dccurred under the point of application of the load and was
generally in the form of a V-shaped group of cracks.

1 Each of the values given in this table is the average of two tests.

808 AMERICAN CERAMIC SOCIETY.

A study of Tabic 2 shows thai the treated samples of a ^i \ <-n
kind and thickness were about 15 per cenl weaker than the
plain u';i^s when the celluloid side was in compression and about
40 per cent Stronger when the celluloid side was in tension

Summary.

The following recapitulation may be made of the salient points
of these tests:

1. It is possible to determine some of the physical properties
of sheet glass in a fairly accurate manner by the use of a com
paratively simple testing apparatus.

2. In both the impact and cross bend tests, the blown window
glass was stronger than the plate glass.

3. While the energy of impact necessary to produce fracture
increased with the thickness of the plate glass, the amount of
cracking occurring at the time of initial fracture also increased.

4. The force of the blow required to crack a piece of glass
of a given kind and thickness was not materially increased by the
celluloid coating, but the shattered glass remained in place wdhle
the plain glass was scattered.

5. In the cross bend tests the treated glass of a given kind and
thickness was about 15 per cent weaker than the plain glass
when the celluloid coated side was in compression and about
40 per cent stronger when the celluloid coated side was in tension.

American Ceramic Society

YEAR BOOK
1918

'<

CONTENTS.

Page.
Officers for the Year I9i8-'i9 813

Officers of Local Sections, 1918-' 19 814

Committees, 1918 814

Honorary Members 816

Active Members 816

Associate Members 820

In Memoriam 849

Contributing Members 850

Honor Roll — -Members 853

Honor Roll — Student Branches 854

Rules of the Society 855

Publications 865

Annual Report of the Board of Trustees 866

Reports of Committees and Student Branches 868

In Retrospect 876

List of Technical Papers in Volume 1 878

Index to Volume I 883

OFFICERS.

For the Year 1918-1919.

President.
Homer F. Staley Washington, D. C.

Vice-President.
A. F. Greaves-Walker Baltimore, Md.

Secretary.
Charles F. Binns Alfred, New York

Treasurer.
R. K. Hursh x Urbana, Illinois

TRUSTEES.
The four above named officers and

Adolph E. Hotting er, term expires 191 9.
Earl T. Montgomery, term expires 1920.
R. D. Landrum, term expires 1921.
George H. Brown, Retiring President.

EX-PRESIDENTS AND WHERE MEETINGS HAVE BEEN HELD.

Herbert A. Wheeler 1 899-1 900 Columbus, Ohio

Karl Langenbeck 1900-1901 Detroit, Mich.

Charles F. Binns 1901-1902 Old Point Comfort, Va.

Ernest Mayer 1902-1903 Cleveland, Ohio

Edward C. Stover 1903-1904 Boston, Mass.

Francis W. Walker 1 904-1 905 Cincinnati, Ohio

William D. Gates 1905-1906 Birmingham, Ala.

Willard D. Richardson • 1 906-1 907 Philadelphia, Pa.

Stanley G. Burt 1907-1908 St. Louis, Mo.

Albert V. Bleininger 1908-1909 Columbus, Ohio

Ross C. Purdy 1909-1910 Pittsburgh, Pa.

H. Ries 1910-191 1 Trenton, N. J.

Charles Weelans 191 1-1912 Chicago, 111.

Arthur S. Watts 1912-1913 Washington, D. C.

Ellis Lovejoy 1913-1914 Wheeling, W. Va.

Cullen W. Parmelee 1914-1915 Detroit, Mich.

Richard R. Hice 1915-1916 Cleveland, Ohio

Lawrence E. Barringer 1916-1917 New York, N. Y.

Seorge H. Brown 1917-1918 Indianapolis, Ind.

814 V >URNAL OF THE

OFFICERS OF LOCAL SECTIONS, 1918.

Northern Ohio Section.

E. P. Poste, Chairman Bryan A. Rice, Secretary

R. D. Landrum, Councilor

St. Louis Section.
H. A. Wheeler, Chairman Gail R. Truman, Secretary

D. T. Farnham, Councilor

New England Section.

W. H. Greuby, Chairman C. J. Hudson, Secretary pro tent

C. H. Kerr, Councilor

Central Ohio Section.

F. K. Pence, Chairman John M. Morton, Secretary

A. S. Watts, Councilor

Chicago Section.
W. D. Gates, Chairman Fred B. Ortman, Secretary

C. W. Parmelee, Councilor

New York- State Section.
L. E. Barringer, Chairman J. B. Shaw, Secretary

C. F. Binns, Councilor

Pittsburgh District Section.
R. R. Hice, Chairman F. H. Riddle, Secretary

F. W. Walker, Sr., Councilor

COMMITTEES, 1918.

Committee on Publications.
L. E. Barringer, Chairman H. Ries

G. H. Brown, Editor A. V. Bleininger

E. W. Tillotson

Committee on War Service.
A. F. Greaves-Walker, Chairman R. H. Minton

A. V. Bleininger C. F. Binns

L. E. Barringer E. W. Washburn

A. S. Watts R. T. Stull

C. H. Kerr R. D. Landrum

AMERICAN CERAMIC SOCIETY.

815

Committee on Membership.

R. C. Purdy, Chairman R. H. Minton

R. D. Hatton
F. H. Riddle
F. K. Pence
J. B. Shaw

R. D. Landrum
R. L. Clare
F. B. Ortman
T. H. Sant

Committee on Standards.

M. F. Beecher, Chairman
C. H. Kerr
J. B. Shaw
E. P. Poste

E. E. F. Creighton
C. W. Berry

F. H. Riddle
R. M. Howe

C. F. Binns
H. Wilson
R. K. Hursh

F. A. Kirkpa trick
C. B. Harrop

G. H. Brown
R. T. Stull

Committee on Cooperation.

C. H. Kerr, Chairman

G. E. Barton

W. E. Emley

J. P. B. Fiske

W. D. Gates

C. T. Harris

Committee on Rules.

C. W. Parmelee, Chairman
A. S. Watts
F. H. Riddle

E. Hart

F. R. Henry
D. F. Riess
K. Seaver

C. L. Sebring

F. W. Walker, Sr.

S. G. Burt

A. F. Greaves-Walker

Committee on Local and Student Sections.

R. R. Hice, Chairman
W. D. Gates
M. E. Gregory

F. K. Pence
H. A. Wheeler

Committee on Chemical Exposition.

L. E. Barringer, Chairman R. D. Landrum

C. Purdy

R

C. H. Kerr
K. Seaver
C. F. Binns
P. H. Swalm
F. B. Ortman

H. Schmidt
A. Staudt
H. Ries
R. T. Stull
R. H. Minton

816 JOURNAL OF THE

Committee on Papers and Programs.

R. H. Minton, Chairman A. E. Williams

A. A. Klein C. E. Jackson

R. T. Stull W. E. Emley

P. H. Bates E. T. Montgomery

E. P. Poste D. T. Farnham

Committee on Summer Session.

F. K. Pence, Chairman E. Ogden

C. W. Berry R. D. Landrum

C. D. Fraunfelter J. W. Sanders

HONORARY MEMBERS.

English Ceramic Society, Mr. R. L. Johnson, President, Oulton Rocks, Stone,
Staffordshire, England.

Storer, Mrs. Maria Longworth (Mrs. Bellamy Storer), Founder of the Rook-
wood Pottery Company, Cincinnati, Ohio.

Branner, John C, Ph.D., President, Leland Stanford Junior University.

ACTIVE MEMBERS.

Arnold, H. C, Urbana, 111., Honor Roll.

Aubrey, Arthur J., Youngstown, Ohio, Assistant General Superintendent,
Bessemer Limestone Company.

Babcock, Grover, Pittsburgh, Pa., Mellon Institute.

Barringer, L. E., Schenectady, New York, Engineer of Insulations, General
Electric Company.

Bates, P. H., Pittsburgh, Pa., Chemist, Bureau of Standards.

Beecher, Milton F., Worcester, Mass., Refractory Engineer, Norton Com-
pany.

AMERICAN CERAMIC SOCIETY. 817

Binns, Charles F., Alfred, New York, Director, New York State School Clay-
working and Ceramics.

Blair, Marion W., Thornton, West Virginia, Superintendent, Thornton Fire
Brick Company.

Blair, William P., 824 B. of L. E. Bldg., Cleveland, Ohio, Secretary, National
Paving Brick Manufacturers' Association.

Bleininger, A. V., Pittsburgh, Pa., Bureau of Standards.

Bloomfield, Charles A., Box 652, Metuchen, New Jersey, Treasurer, Bloom-
field Clay Company.

Boeck, Percy A., 11 Broadway, New York City, Celite Products Company.

Brockbank, Clarence J., Niagara Falls, N. Y., Honor Roll.

Brown, George H., New Brunswick, New Jersey, Director, Department of
Ceramics, Rutgers College.

Burt, S. G., 2349 Ashland Ave., Cincinnati, Ohio, Rookwood Pottery Com-
pany.

Cannan, Wm., Jr., 522 Allen St., Syracuse, New York, Onondaga Pottery.

Clare, Robert L., Woodbridge, New Jersey, Superintendent, Federal Terra
Cotta Company.

Cowan, R. Guy, Cleveland, Ohio, Honor Roll.

Creighton, E. E. F., 27 Wendell Ave., Schenectady, New York, Consulting
Electrical Engineer, General Electric Company.

Emley, Warren E., 3705 Keokuk St., Chevy Chase, D. C, Bureau of Stand-
ards.

Farnham, Dwight T., 3rd National Bank Building, St. Louis, Mo., Industrial
Engineer.

Fickes, Walter M., 1882 Grantham Ave., S. E., Collingwood Station, Cleve-
land, Ohio, The Aluminum Rolling Mill Company.

Fiske, J. Parker B., Arena Bldg., New York City, Fiske and Company, Inc.

Frink, R. L., Lancaster, Ohio, President, Frink Laboratories.

Fulton, C. E., Creighton, Pa., Pittsburgh Plate Glass Company.

Galpin, S. L., 620V2 Sherman Ave., Hutchinson, Kansas.

Garve, T. W., 320 — 17th Ave., Milwaukee, Wisconsin.

Gates, Ellis D., Peoples Gas Bldg., Chicago, Illinois, American Terra Cotta
and Ceramic Company.

Gates, William D., Peoples Gas Bldg., Chicago, Illinois, President and General
Manager, American Terra Cotta and Ceramic Company.

Geijsbeek, Samuel, 604 Blake McFall Bldg., Portland, Oregon, Consulting
Chemical Engineer, Geijsbeek Engineering Company.

Gelstharp, Frederick, Creighton, Pa., Chemist, Pittsburgh Plate Glass Com-
pany.

Grady, Robert F., St. Louis, Mo., Manager, St. Louis Terra Cotta Company.

Greaves-Walker, A. F., Tudor Hall, University Parkway, Baltimore, Md.,
Representative, American Refractories Company.

Harrop, Carl B., Lord Hall, Ohio State University, Columbus, Ohio, Assistant
Professor Ceramic Engineering.

Si8 JOURNAL OF THE

Hart, Edward, Ph.D., Kaston, Pa , Professor Chemistry, Lafayette College,

Henderson, H. B., [538 North IIjk'.i St., Columbus, Ohio, Superintendent.
Orton Pyrometric Cone Company.

Hice, Richard R., Sc.D., Beaver, Pa., State Geologist, Topographic and Geo-
logical Survey Commission of Pennsylvania.

Hill, Ercell C, 401 Vernon Ave., Long Island City, New York, Chemist,
New York Architectural Terra Cotta Company.

Hope, Herford, Beaver Falls, Pa., Honor Roll.

Horning, Roy A., Beaver Falls, Pa., Honor Roll.

Hottinger, A. F., Chicago, Illinois, Northwestern Terra Cotta Company.

Howat, Walter L., Perth Amboy, N. J., Honor Roll.

Howe, Raymond M., Pittsburgh, Pa., Mellon Institute.

Humphrey, Dwight E., 103 Fourth St., Cuyahoga Falls, Ohio, Secretary,
Portage Engineering Company.

Humphrey, Harold P., Washington, New Jersey, Washington Porcelain
Company.

Hursh, Ralph K., Urbana, Illinois, Ceramic Department, University of
Illinois.

Jackson, C. Edward, Wheeling, West Virginia, President, Warwick China
Company.

Jones, Robert W., 62 Dudley Ave., Tompkinsville, N. Y., State Inspector
of Mines and Tunnels.

Karzen, Samuel C, Uhrichsville, Ohio, American Sewer Pipe Company.

Kerr, Charles H., Southbridge, Mass., Director Research Laboratory, Ameri-
can Optical Company.

Kirk, Charles J., New Castle, Pa., President and General Manager, Universal
Sanitary Manufacturing Company.

Kirkpatrick, F. A., Pittsburgh, Pa., Bureau of Standards.

Klein, A. Albert, Worcester, Mass., Research Laboratory, Norton Company.

Klinefelter, T. A., Box 519, Derry, Pa., Ceramic Engineer, Research De-
partment, Westinghouse Electric & Manufacturing Company.

Krehbiel, J. F., 1538 North High St., Columbus, Ohio, Orton Pyrometric
Cone Factory.

Landrum, R. D., Cleveland, Ohio, Manager Service Department, Harshaw,
Fuller and Goodwin Company.

*Langenbeck, Karl, 1625 Hobart St., N. W., Washington, D. C.

Loomis, George A., Corning, New York, Corning Glass Works.

Love joy, Ellis, Schultz Bldg., Columbus, Ohio, Manager, Lovejoy Engineer-
ing Company.

Mayer, A. E., Beaver Falls, Pa., Assistant Secretary, Mayer Pottery Com-
pany.

Minton, R. H., Metuchen, New Jersey, Superintendent, General Ceramics
Company.

* Life member.

AMERICAN CERAMIC SOCIETY. 819

Montgomery, E. T., Detroit, Michigan, Research Manager, Jeffery-Dewitt

Company.
Montgomery, Robert J., Rochester, New York, Bausch and Lomb Optical

Company.
Moore, Joseph K., Cleveland, Ohio, Honor Roll.

Morris, George D., St. Louis, Mo., Evans and Howard Fire Brick Company.
Ogden, Ellsworth, Chelsea Road, Upper Arlington, Columbus, Ohio, Assistant

of Research in Ceramic Engineering, Ohio State University.
Orton, Edward, Jr., Lt.-Col., Washington, D. C, Honor Roll.
Parker, Lemon, 3314 Morganford Road, St. Louis, Mo., Vice-President and

Superintendent, Parker-Russell Mining and Manufacturing Company.
Parmelee, C. W., Urbana, Illinois, Ceramic Engineering Department, Uni-
versity of Illinois.
Pence, F. K., Zanesville, Ohio, Ceramic Engineer, American Encaustic

Tiling Company.
Plusch, H. A., Philadelphia, Pa., Research Engineer, Abrasive Company.
Poste, Emerson P., Elyria, Ohio, Chemical Engineer, Elyria Enameled

Products Company.
Potts, Amos, Mason City, Iowa, Mason City Brick and Tile Company.
Purdy, Ross C, Worcester, Mass., Research Engineer, Norton Company.
Radcliffe, B. S., 15 15 Lumber Exchange Bldg., Chicago, Illinois, Midland

Terra Cotta Company.
*Randall, T. A., Indianapolis, Indiana, Editor, "The Clayworker."
Richardson, W. D., 388 N. 8th Ave., Columbus, Ohio.

Riddle, F. H., Pittsburgh, Pa., Associate Ceramic Chemist, Bureau of Stand-
ards.
Ries, H., Ph.D., Ithaca, New York, Professor of Economic Geology, Cornell

University.
Schurecht, H. G., Lord Hall, Columbus, Ohio, U. S. Bureau of Mines, Ohio

State University.
Shaw, Joseph B., Alfred, New York, Assistant Director, New York State

School of Clayworking and Ceramics.
Silverman, Alexander, Pittsburgh, Pa., Head, Department of Chemistry,

University of Pittsburgh.
Simcoe, George, Box 108, Metuchen, New Jersey.
Spurrier, Harry, 479 Crane Ave., Detroit, Michigan, Chemist, Jeffery-Dewitt

Company.
Staley, H. F., Washington, D. C, Bureau of Standards, Division VIII.
Stephani, William J., Crum Lynne, Pa., Superintendent, O. W. Ketcham

Terra Cotta Company.
Stout, Wilbur, Columbus, Ohio, Ohio State University.
Stover, Edward C, 474 West State St., Trenton, New Jersey, Assistant

General Manager, Trenton Potteries Company.

* Life member.

820 JOURNAL OF THE

Stull, Ray T., Lord Hall, Columbus, Ohio, Chief Engineer, Ceramic Re-
search Station, U. S. Bureau Of Mines.

Teetor, Paul, Detroit, Michigan, Ceramic Engineer, Jeffery Dewitt Company.

Tillotson, E. W., Ph.D., Pittsburgh, Pa., Mellon Institute.

Walker, Francis W., Beaver Falls, Pa., Secretary and Treasurer, Art Tile
Company.

Walker, Francis W., Jr., Heaver Palls, Pa., Honor Roll.

Ward, J. W., 628 Main St., I,atrohe, Pa., General Superintendent, Pitts-
burgh High Voltage Insulator Company.

Washburn, E. W., Ph.D., TJrbana, Illinois, Professor Ceramic Engineering,
University of Illinois.

Watts, Arthur S., Lord Hall, Columbus, Ohio, Professor Ceramic Engineering,
Ohio State University.

Wheeler, Herbert A., 408 Locust St., St. Louis, Mo., Consulting Engineer.

Whitford, W. G., Chicago, 111., Honor Roll,

Williams, Arthur E., Pittsburgh, Pa., Bureau of Standards.

Williams, Ira A., Seattle, Washington, Ceramist, Mines Station, U. S. Bureau
of Mines, University of Washington.

FOREIGN ACTIVE MEMBERS.

Cronquist, Gustaf Werneisson, Rikstel 153, Helsingborg, Sweden.

Davis, N. B., 410 Union Bank Bldg., Ottawa, Canada, M. J. O'Brien, Ltd.

Keele, Joseph, Ottawa, Canada, Department of Mines, Mines Branch.

Knote, J. M., Sault Ste. Marie, Ontario, Canada, Mines Department, L. S.
Corporation.

Takahashi, Kakuichiro, Hodogaya, Near Yokohama, Japan, General Super-
intendent, Nippon Garas Kogyo Kaisha.

Worcester, Wolsey, 1722 — -25 A St., W., Calgary, Alberta, Canada.

ASSOCIATE MEMBERS.

Abrams, Duff A., Chicago, Illinois, In Charge of Structural Materials Re-
search Laboratory, Lewis Institute.

Acheson, Edward G., Niagara Falls, New York, President, The Acheson
Company.

Ahlman, Louis F., 720 Electric Bldg., Cleveland, Ohio, Salesman, Harshaw,
Fuller and Goodwin Company.

Allen, Edwin M., 208 S. LaSalle St., Chicago, Illinois, American Refractories
Company.

Allen, F. B., Woodbridge,|New Jersey, Honor Roll.

AMERICAN CERAMIC SOCIETY. 821

Anderson, George O., Parkersburg, West Virginia, Secretary and Treasurer,
General Porcelain Company.

Anderson, Robert E., Worcester, Mass., Honor Roll.

Applegate, D. H., Jr., Red Bank, New Jersey, Manager, Crescent Brick Com-
pany.

Arbenz, Fred J. A., Chillicothe, Ohio, Manager, Florentine Pottery.

Arbogast, F. J., 154 Belmont Ave., Crafton, Pa., Salesman, Hachmeister
Lind Chemical Company.

Arbogust, Charles O., 808 Lakeside Place, Chicago, Illinois, Erwin and Wasey.

Ashbaugh, Charles C, East Liverpool, Ohio, Secretary and Treasurer, West
End Pottery Company.

Austin, Arthur O., 326 North 6th St., Barberton, Ohio, Chief Engineer, Ohio
Insulator Company.

Austin, James L., 37 Banks St., Waltham, Mass., Ceramic Engineer, Waltham
Grinding Wheel Company.

Avars, Erling E., 2307 Elsinor Ave., Baltimore, Md., American Refractories
Company.

Ayer, Kenneth R., Honor Roll.

Ayers, Elwood B., 7th St. and Tabor Road, Philadelphia, Pa., Philadelphia
Textile Machinery Company.

Baccus, Mrs. Julia F., Cleveland, Ohio, Enameling Expert, Cleveland Metal
Products Company.

Bacon, Charles C, Tacony, Philadelphia, Pa., Secretary, Ross-Tacony
Crucible Company.

Bacon, Raymond F., Pittsburgh, Pa., Director, Mellon Institute, Honor Roll.

Baggs, Arthur E., Marblehead, Mass., Manager, Marblehead Pottery Com-
pany.

Bailar, John C, Golden, Colorado, Assistant Professor of Chemistry, Colorado
School of Mines.

Bainbridge, Walter L., Newark, Ohio, Chemical Engineer, The American
Bottle Company.

Baker, G. V., Barnard, New York, Superintendent, Pennsylvania Feldspar
Company.

Balmert, Richard M., 1825 Franklin Ave., Portsmouth, Ohio.

Balz, George A., Perth Amboy, New Jersey, Manager, Didier-March Com-
pany.

Bardush, J. F., Grand Rapids, Michigan, Grand Rapids Refrigerator Com-
pany.

Barker, Charles E., Matawan, New Jersey, Atlantic Tile Manufacturing
Company.

Bartells, H. H., Bessemer, Pa., Honor Roll.

Bartlett, Edward E., Sapulpa, Oklahoma, Bartlett-Collins Glass Company.

Barton, G. E., 227 Pine St., Millville, New Jersey, Chief Chemist and Technical
Expert, Whitall Tatum Company.

Bassett, Leon B., Worcester, Mass., Ceramic Engineer, Norton Company.

J( lURNAL OF THE

Bates, Charles E., (ialesburg, Illinois, Honor Roll

Bedson, William, Lawrence Road, Trenton, New Jersey, Assistant Man
Elite Pottery Company.

Bell, Fred S., 1318 Wright Bldg., St. Louis, Mo., Vice-President, Mandl<
Clay Mining Company.

Bell, M. Llewellyn, 901 South Ave., Wilkinsburg, Pa., Carnegie Steel Com
pany, Braddock, Pa.

Bellamy, Harry T., Chicago, Illinois, Hawthorn Plant, Western Electric Com-
pany.

Benner, William J., Hebron, North Dakota, Honor Roll.

Bently, Louis L., Beaver Falls, Pa., Superintendent, Armstrong Cork Com-
pany.

Berger, William S., 1601 Woodburn Ave., Covington, Kentucky, Secretary,
Cambridge Tile Manufacturing Company.

Bergmans, Carl, 263 West Main St., Zanesville, Ohio, Ceramist, American
Encaustic Tiling Company.

Berry, Charles W., St. Louis, Mo., Chemist, Laclede-Christy Clay Products
Company.

Best, Harold A., Ottawa, Illinois, Honor Roll.

Bish, H. A., Findlay, Ohio.

Bissell, G. F., Ottawa, Illinois, Chicago Retort and Fire Brick Company.

Blackmer, Edward L., St. Louis, Mo., Assistant Superintendent, Blackmer
and Post Pipe Company.

Blake, Alfred E., Pittsburgh, Pa., Fellow and Assistant Professor, Mellon
Institute.

Blewett, J. B., Wellsville, Ohio, McLain Fire Brick Company.

Blinks, Walter M., Kalamazoo, Michigan, Manager, Michigan Enameling
Works.

Bole, G. A., Alfred, New York, Professor of Chemistry, Alfred University.

Booraem, J. Francis, 52 Vanderbilt Ave., New York City, Treasurer and
Manager, American Enameled Brick and Tile Company.

Booze, M. C, Columbus, Ohio, U. S. Bureau of Mines.

Bosworth, Harold O., Denver, Colorado, President, Denver Fire Clay Com-
pany.

Boughey, Joseph, 584 Roosevelt Ave., Trenton, New Jersey, Manager,
Porcelain Department, Chas. Engelhard, Newark, New Jersey.

Bowman, Oliver O., 2nd, Trenton, New Jersey, Ceramist, Trenton Fire Clay
and Porcelain Company.

Bowman, William J. J., Trenton, New Jersey, Trenton Fire Clay and Por-
celain Company.

Bowne, Edward, Cloverport, Kentucky, Murray Roofing Tile Company.

Bradshaw, A. R., Minerva, Ohio, Owen China Company.

Bragdon, William V., 2336 San Pablo Ave., Berkeley, California, "The Tile
Shop."

AMERICAN CERAMIC SOCIETY. 823

Brain, George, Newcastle, Pa., Manager, Universal Sanitary Manufacturing
Company.

Bramlett, Julian T., R. F. D. No. 1, Enid, Mississippi, General Manager,
Bramlett Clay Property.

Brand, J. J. Fred, Box 472, Roseville, Ohio.

Brandt, George F., 458 Broadway, Bedford, Ohio, Superintendent, Bedford
China Company.

Bray, Archie C, Helena, Montana, Assistant Superintendent, Western Clay
Manufacturing Company.

Brewster, Robert, 2814 Cambridge Ave., Chicago, Illinois, Dental Porcelain
Manufacturing Company.

Breyer, Frank G., Palmerton, Pa., Chief of Research Department, New
Jersey Zinc Company.

Briggs, Leonard S., South River, New Jersey, National Fireproofing Company.

Brockmann, Edward A., 192 West Clark St., Chicago, Illinois, Salesman,
Roessler and Hasslacher Chemical Company.

Broga, Wilson C, Worcester, Mass., Honor Roll.

Brogdon, J. S., 70 l /a Peachtree St., Atlanta, Georgia, Consulting and Manu-
facturing Chemist.

Brower, Fred W., Uhrichsville, Ohio, Belden Brick Company.

Brown, Davis, Bucyrus, Ohio, American Clay Machinery Company.

Brown, Edmund, Perrysburg, Ohio.

Brown, L. K., Zanesville, Ohio, Vice-President, The Burton Townsend Com-
pany.

Brown, Philip King, M.D., 925 Hyde St., San Francisco, California.

Brown, R. E., Utica, Illinois, Utica Fire Brick and Clay Company.

Brown, Richard P., Wayne Junction, Philadelphia, Pa., President, Brown
Instrument Company.

Brown, Wilbur F., Muncie, Indiana, Honor Roll.

Browning, T. S., Vanport, Pa., McLain Fire Brick Company.

Brownlee, William K., Toledo, Ohio, President and General Manager, The
Buckeye Clay Pot Company.

Bruner, Willard L., 196 Water St., Perth Amboy, New Jersey, Ceramic
Chemist, Roessler and Hasslacher Chemical Company.

Brush, George S., 11 12 Lexington Ave., Zanesville, Ohio, General Manager,
The Brush-McCoy Pottery Company.

Bryan, M. L., Seattle, Washington, Superintendent Terra Cotta Depart-
ment, Denny Renton Clay and Coal Company.

Bucher, E. G., 203 Sherman Ave., Pontiac, Illinois.

Bucher, L. R., Vandalia, Mo., Honor Roll.

Burns, Ray A., St. Louis, Mo., Engineer of Sales Department, Laclede-
Christy Clay Products Company.

Burroughs, Francis H., Trenton, New Jersey, Assistant Chemist, Star Por-
celain Company.

Burton, Charles G., Peru, Indiana, Manager, Peru Electric Company.

824 l< >URNAL OF THE

Butterworth, Frank W., Danville, Illinois. General Manager, Western Brick
Company

Byers, Louis L., Philadelphia, Pa., Honor Roll.

Cable, Davis A., Fast Sparta, Ohio, General Superintendent, U. S. Roofing
Tile Company.

Cable, Miss Margaret K., Grand Forks, North Dakota, Instructor in Ceramics,
University of North Dakota.

Callaghan, J. P., Sharon, Pa., Superintendent, Sharon Clay Products Com-
pany.

Campbell, A. R., Middlesex Ave., Metuchen, New Jersey, Federal Terra Cotta
Company.

Carder, Frederic R., Corning, New York, Steuben Glass Works.

Carman, C. F., Berkeley Springs, West Virginia, President, National Silica
Works.

Carrier, Augustus, Detroit, Michigan, Chemist, Detroit Dental Manu-
facturing Company.

Carter, Benjamin F., Peoria, Illinois, Honor Roll.

Carter, C. E., 119 Bridge St., Peoria, Illinois, General Manager, Carter Brick
Company.

Casey, Charles L., Cambridge, Ohio, President, Guernsey Earthenware
Company.

Casey, J. B., Colfax, Indiana, Vice-President and General Manager, Colfax
Drain Tile Company.

Cassady, Bertram L., 614 Aten Ave., Wellsville, Ohio, Superintendent, Mc-
Lain Fire Brick Company.

Cermak, Frank, 116 4th Ave., Schenectady, New York, Foreman, Porcelain
Works, General Electric Company.

Chambers, A. R., 466 Greenwood Ave., Trenton, New Jersey, Manufacturer
of Fire Brick.

Chandler, A. H., 1618 Frick Bldg., Pittsburgh, Pa., Manager, Refractories
Department, Pittsburgh Plate Glass Company.

Chapin, F. H., 900 Schofield Bldg., Cleveland, Ohio, Manager, Hydraulic
Press Brick Company.

Chapman, Miss Dorothy P., 31 Shattuck St., Worcester, Mass., Ceramic
Division, Norton Company.

Chase, John A., Honor Roll.

Chatham, C. W., 1407 Arrott Bldg., Pittsburgh, Pa., Manager, Eagle-Picher
Lead Company.

Cheney, M. B., Briggsdale, Ohio.

Child, J. L., Findlay, Ohio, Hancock Brick and Tile Company.

Chinery, Frank L., Cincinnati, Ohio, Eagle-Picher Lead Company.

Chormann, O. L, Major, Niagara Falls, New York, Honor Roll.

Christie, J. M., 1950 East 90th St., Cleveland, Ohio, Salesman, Roessler and
Hasslacher Chemical Company.

Christman, C. E., Pittsburgh, Pa., President, Federal Enameling and Stamp-
ing Company.

AMERICAN CERAMIC SOCIETY. 825

Clark, A. B., Stanford University, California, Head, Department of Graphic

Arts, Leland Stanford Junior University.
Clark, Harold E. (Address unknown.)
Clark, John, 292 Lockwood St., Astoria, New York, Superintendent, New

York Architectural Terra Cotta Company.
Clark, William M., E. 152nd St., Cleveland, Ohio, Manager, Euclid Glass

Division, National Lamp Works.
Clark, Willis H., Midland, Michigan, Dow Chemical Company.
Claudon, Charles F., Ottawa, Illinois, Superintendent, Federal Plate Glass

Company.
Clayter, Frederic C, Pittsburgh, Pa., Honor Roll.
Cliff, Ray T., Sebring, Ohio, The Saxon China Company.
Cole, M. J., Logan, Ohio, Superintendent and Treasurer, The Logan Clay

Products Company.
Conkling, Samuel O., Philadelphia, Pa., Superintendent, Conkling-Arm-

strong Terra Cotta Company.
Cook, Charles H., Trenton, New Jersey, Cook Pottery Company.
Cook, John H., Minerva, Ohio, The Owen China Company.
Cooke, May E., 1550 Clifton Ave., Columbus, Ohio.
Cooke, Raymond D., Terre Haute, Indiana, Chemist, Columbian Enameling

and Stamping Company.
Coors, H. F., Golden, Colorado, Manager, Herold China and Pottery Com-
pany.
CoVan, H. E., 1961 East 116th St., Cleveland, Ohio, Price Electric Pyrometer

Company.
Cox, Harold N., Silver Lake, New Jersey, Thomas A. Edison Interests.
Cox, S. Frank, Mt. Vernon, Ohio, Research Chemist, Pittsburgh Plate Glass

Company.
Crane, Charles W., 17 Battery Place, New York City, Mining Fire Clays and

Silica Sand.
Crawford, Edward A., Canastota, R. D., New York.
Crawford, J. L., 31 Watsessing Ave., Bloomfield, New Jersey.
Creusere, G. C, Perrysville, Ohio, Manager, Mohawk Clay Products Com-
pany.
Crew, H. F., Zanesville, Ohio, Honor Roll.
Cronin, W. Kress, 320 West 5th St., East Liverpool, Ohio, Standard Pottery

Company.
Cunningham, Matthias F., Waltham, Mass., Superior Corundum Wheel

Company.
Cutting, J. W., Cambridge, Mass, Superintendent, A. H. Hews Company.
Dailey, Ernest W., Utica, Mo., General Superintendent, Shale Hill Brick and

Tile Company.
Danielson, Ralph R., Des Plaines, Illinois, Ceramic Chemist, Benjamin

Electric Manufacturing Company.
Darling, James, Box 300, Kenova, West Virginia.

826 JOURNAL OF THE

Darlington, Homer T., Box 7.56, Natrona, Pa., Metallurgical Enjineer,

Pennsylvania Salt Manufacturing Companj
Davenport, R. W., Detroit, Michigan, Jeffery Dewitt Company
Davis, Harold E., 115 Macon St., Waltham, Mass, Superintendent, Waltham

Grinding Wheel Company.
Davis, John B., D.D.S., 22 Bast Greenwood Ave ., Lansdowne, Pa.
Davis, John H., 792 Beacon St., Boston, Mass.
Day, Arthur L., Ph.D., Sc.D., Washington, I) C, Director, Geophysical

Laboratory.
Dean, Charles A., 79-1 2th Ave, Columbus, Ohio.
Deaver, L. A., 29 Dodge Ave., Akron, Ohio.

Dell, Jack M., 6100 Manchester Ave., St. Louis, Mo., Assistant Super-
intendent, Missouri Fire Brick Company.
DeVoe, Charles H., Old Bridge, New Jersey, Superintendent, Old Bridge

Enameled Brick and Tile Company.
DeVol, E. M., Zanesville, Ohio, Superintendent, Zanesville Stoneware Com-
pany.
Dickey, Fred L., Kansas City, Mo., General Manager, W. S. Dickey Clay

Manufacturing Company.
Dinsmore, B. B., Trenton, New Jersey, General Manager, Imperial Porcelain

Works.
Dittmar, Carl, 518 Union Central Bldg., Cincinnati, Ohio, Resident Manager,

Roessler and Hasslacher Chemical Company.
Dobbins, T. Monroe, Camden, New Jersey, Secretary and Treasurer, Camden,

Pottery Company.
Dolley, Charles S., Fort Madison, Iowa, Chemical Engineer, Keramoid

Manufacturing Company.
Dornbach, William E., Box 190, Baltimore, Md., American Refractories

Company.
Dorst, Max, Honor Roll.
Drakenfeld, B. F., Jr., 50 Murray St., New York City, Treasurer, B. F.

Drakenfeld and Company, Inc.
Dressier, Philip D'H., Box 516, Zanesville, Ohio.

Dufour, George F., Y. M. C. A., Hamilton, Ohio, Consulting Engineer.
Dunn, Frank B., Conneaut, Ohio, President and Treasurer, Dunn Wire Cut

Lug Brick Company.
Durant, E. M., 825 East 7th St., Los Angeles, California, President, Pacific

Sewer Pipe Company.
Duval, A. L., Washington, Pa., Chief Chemist, Hazel-Atlas Glass Company.
Earle, Robert W., Canal Dover, Ohio, Vice-President, Dover Manufacturing

Company.
Ebinger, D. H., Jr., Columbus, Ohio, D. A. Ebinger Sanitary Manufacturing

Company.
Eckert, E. W., Box 65, Derry, Pa., Electrical Engineer, Pittsburgh High

Voltage Insulator Company.

AMERICAN CERAMIC SOCIETY. 827

Edgar, David R., Metuchen, New Jersey, Assistant General Manager, Edgar
Brothers Company.

Ellerbeck, William L., D.D.S., 703 Boston Bldg., Salt Lake City, Utah,
President and Manager, Nephi Plaster and Manufacturing Company.

Emrick, W. A., Derry, Pa., Assistant to the General Superintendent, Pitts-
burgh High Voltage Insulator Company.

Eskesen, Bermet K., Matawan, New Jersey, Superintendent, New Jersey
Mosaic Tile Company.

Eskesen, E. V., 149 Broadway, New York City, General Manager, New
Jersey Terra Cotta Company.

Evatt, Frank G., 705 Fulton Bldg., Pittsburgh, Pa., District Manager, Atlantic
Terra Cotta Company.

Fackt, George P., Denver, Colorado, General Manager, Denver Terra Cotta
Company.

Farren, Miss Mabel C, 3600 Forbes St., Pittsburgh, Pa., Private Studio.

Fenton, Harry W., Mogadore, Ohio, Assistant General Manager, Akron,
Smoking Pipe Company.

Ferguson, Richard D., Lincoln, Nebraska, Honor Roll.

Ferguson, Robert F., Pittsburgh, Pa., Honor Roll.

Fettke, Charles R., Pittsburgh, Pa., Carnegie Institute of Technology.

Fisher, Douglas J., Sayreville, New Jersey, Sayre and Fisher Company.

Fisher, George P., Ottawa, Illinois, National Fireproofing Company.

Fitzpatrick, John, 533 10th St., Niagara Falls, New York, Carborundum
Company.

Foersterling, Hans, Ph.D., Jamesburg, New Jersey, Second Vice-President,
Roessler and Hasslacher Chemical Company.

Footitt, Frank F., Room 216, Ceramic Building, Urbana, Illinois, University
of Illinois.

Ford, G. Bergen, Caudler Bldg., Atlanta, Ga., B. Mifflin Hood Company.

Forester, Herbert, Box 37, Cleveland, Ohio, Principal, Veritas Firing System
Company.

Forman, L. P., Arnold, Pa., American Window Glass Company.

Forst, Arthur D., Trenton, New Jersey, President, Robertson Art Tile Com-
pany.

Forst, D. Parry, 73 North Clinton Ave., Trenton, New Jersey, Robertson
Art Tile Company.

Fowler, William E., Paxtonville, Pa., Paxton Brick Company.

Franzheim, C. Merts, Wheeling, West Virginia, The Charles M. Franzheim
Company.

Fraser, W. B., Dallas, Texas, President, Fraser Brick Company.

Fraunfelter, Charles D., Zanesville, Ohio, President, Ohio Pottery Company.

Freese, H. H., 446 South Union St., Galion, Ohio, Mechanical Engineer,
E. M. Freese Company.

Frerichs, Wm. D., in Broadway, New York City, Federal Terra Cotta Com-
pany.

828 J( HJRNAL OF THE

Frey, Frank A., Milwaukee, Wisconsin, Enameling Department, Gender,

Paeschke and Prey Company.
Fridricksen, Christian, Wheeling, West Virginia, Wheeling Tile Company.
Frink, Urban W., 17 Power House St., South Boston, Mass., Manager, New

England Branch, E. J. Woodiaon Company. ,
Fritz, E. H., Box 510, Derry, Pa., Ceramic Engineer, Westinghouse Electric

.mil Manufacturing Company.
Frost, Leon J., 11700 Shadeland Ave., Cleveland, Ohio, Chemist and Pur-
chasing Agent, Enamel Products Company.
Fuhrmann, F. J. (Address unknown.)
Fuller, Ralph L., 11 19 Guardian Bldg., Cleveland, Ohio, Harshaw, Fuller and

Goodwin Company.
Fulper, William H., Flemington, New Jersey, Secretary-Treasurer, Fulper

Pottery.
Fulton, Kenneth I., Portland, Indiana, Honor Roll.
Fulweiler, Walter H., 319 Arch St., Philadelphia, Pa., Chief Chemist, United

Gas Improvement Company.
Gahris Willard I., Scbring, Ohio, Limoges China Company.
Galloway, Walter B., Philadelphia, Pa., President, Galloway Terra Cotta

Company.
Garrison, Amos, 417 South First St. West, Salt Lake City, Utah.
Gates, A. W., Colchester, Illinois, President and Treasurer, Gates Fire Clay

Company.
Gates, Major E., Terra Cotta, Illinois, Superintendent, American Terra

Cotta and Ceramic Company.
Gavin, Gordon P., Kalamazoo, Michigan, Superintendent, Kalamazoo

.Sanitary Manufacturing Company.
Gaylord, George L., 14 Holland Ave., WTestfield, Mass., General Manager,

Vitrified Wheel Company.
Gehrig, Edward F., Detroit, Michigan, Research Engineer, Detroit Stove

Works.
Geiger, Carl F., 7 Baltimore St., Dayton, Ohio, Chief Draughtsman and

Engineer, The Manufacturers Equipment Company.
Geiger, Charles F., Department of Ceramics, Rutgers College, New Bruns-
wick, N. J.
Geisinger, E. E., Rochester, New York; Chemist, The Pfaudler Company.
Gerber, Albert C, Toledo, Ohio, Honor Roll.
Gibbs, Arthur E., 1006 Widener Bldg., Philadelphia, Pa., Head of Research

Department, Pennsylvania Salt Manufacturing Company.
Giesey, V. A., 318 Citizens Bldg., Cleveland, Ohio, President, National

Refractories Company.
Gilbert, Frank W., Rochester, Pa., Pittsburgh Grinding Wheel Company.
Gildard, Warren R., Ballston Lake, New York, Research Laboratory, General

Electric Company, Schenectady, N. Y.
Gillinder, James, 1 1 Orange St., Port Jervis, New York, Superintendent,

Gillinder Brothers.

AMERICAN CERAMIC SOCIETY. 829

Gilmore, R. B., E. Liberty Branch, Y. M. C. A., Pittsburgh, Pa., Ceramic

Engineer, Vesuvius Crucible Company.
Gladding, Augustus L., Lincoln, California, Gladding, McBean and Company.
Godfrey, Miss Lydia B., Grand View-on-Hudson, New York, Director, Glen

Tor Keramic Studio.
Golding, Charles E., 217 South Warren St., Trenton, New Jersey, Manager,

Trenton Department, Golding Sons Company.
Goldner, G. E., Dayton, Ohio, Constructing Engineer, Manufacturers Equip-
ment Company.
Goldsmith, B. B., 19 East 74th St., New York City, Vice-President, American

Lead Pencil Company.
Good, Harry N., Pittsburgh, Pa., Ceramic Engineer, Pittsburgh Testing

Laboratory.
Gorton, Arthur F., Ph.D., Toledo, Ohio, Research Physicist, Buckeye Clay

Pot Company.
Grafton, Charles O., Muncie, Indiana, Treasurer and General Manager,

Gill Clay Pot Company.
Grainer, John S., Box 21, Spring Lake, Michigan, Challenge Refrigerator

Company.
Grampp, Otto, 205 Farwell Ave., Milwaukee, Wis., Superintendent, A. J.

Lindeman and Hoverson Company.
Grant, DeForest, 101 Park Ave., New York City, President, Federal Terra

Cotta Company.
Grant, W. Henry, St. Marys, Pa., Elk Fire Brick Company.
Green, J. L., St. Louis, Mo., President, Laclede-Christy Clay Products

Company.
Greene, R. W., Mayfield, Kentucky, Kentucky Construction and Improve-
ment Company.
Greener, George C, 198 Clarendon St., Boston, Mass., Director, North

Bennett Street Industrial School.
Greenwood, G. W., Uniontown, Pa., United Fire Brick Company.
Gregori, John N., Chicago, Illinois, Kiln Burner, Northwestern Terra Cotta

Company.
Gregorius, Thomas K., Corning, New York, Ceramic Engineer, Corning Glass

Works.
Gregory, M. Creveling, Corning, New York, Ceramic Engineer, Corning Brick,

Terra Cotta and Tile Company.
Gregory M. E., Corning, N. Y., Proprietor, Corning Brick, Terra Cotta and

Tile Co.
Grueby, William H., Boston, Mass, President and General Manager, Grueby

Faience Company.
Guastavino, Rafael, Jr., Fuller Bldg., Broadway and 23rd St., New York

City, President, R. Guastavino Company.
Gunniss, William H., Black Eagle Club, Great Falls, Montana.
Haaf, George, 1501 Milton Ave., Solvay, New York, General Foreman, Pass

and Seymour, Inc.

830 JOURNAL OF THE

Haeger, D. C, 26a Grant Place. Aurora, Illinois, President. Kaeger Brick and

Tile Company.

Haeger, E. H., Dundee, Illinois, Haeger Prick and Tile Company.

Haley, Mark A., Syracuse, New York, Onondaga Pottery Company.

Hall, Clarence A., t8 McPherson St., Mount Airy, Pa., Research Depart
inent, Pennsylvania Salt Manufacturing Company.

Hall, Herman A., Nelsonville, Ohio, Pulverized Fire Clay Company.

Hall, Robert T., Jvast Liverpool, Ohio, Secretary and Treasurer, Hall China
Company.

Hall, William C, 1302 — 3rd National Bank Building, Atlanta, Georgia, Vice-
President, Atlanta Terra Cotta Company.

Hamilton, James, Trenton, New Jersey, Superintendent, Ideal Pottery of
The Trenton Potteries.

Handke, Paul A., 260 S. Academy St., Galesbury, Illinois, Purington Paving
Brick Company.

Hanna, Harold H., Crystal City, Mo., Pittsburgh Plate Glass Company.

Hansen, Abel, Fords, New Jersey, Proprietor and Manager, Fords Por-
celain Works.

*Hansen, Harry M., Maplehurst, Metuchen, New Jersey.

Harbeck, H. F., Grand Haven, Michigan, President, Challenge Refrigerator
Company.

Hardesty, B. D., Syracuse, New York, Iroquois China Company.

Harding, C. Knox, 6318 Stoney Island Blvd., Chicago, Illinois, Consulting
Engineer.

Hardy, Isaac E., Momence, Illinois, Superintendent, Tiffany Enameled Brick
Company.

Hare, Robert L., Mexico, Mo., A. P. Green Fire Brick Company.

Harker, H. N., East Liverpool, Ohio, General Manager, Harker Pottery
Company.

Harper, John L., Niagara Falls, New York, Hydraulic Power Company.

Harris, Charles T., 356 Leader Building, Cleveland, Ohio, Secretary, Hollow
Building Tile Manufacturers' Association of America.

Harvey, F. A., Syracuse, New York, Laboratory Physicist, Solvay Process
Company.

Hasburg, John W., 11 19 La Salle Ave., Chicago, Illinois, President, John
W. Hasburg Company, Inc.

Hasslacher, Jacob, 100 William St., New York City, President, Roessler and
Hasslacher Chemical Company.

Hastings, Francis N., Hartford, Conn., Hartford Faience Company.

Hatton, Richard D., 1673 Railway Exchange Bldg., St. Louis, Mo., Vice-
President and General Manager, Laclede- Christy Clay Products Company.

Heilman, Karl J., Tiffin, Ohio, Proprietor, Tiffin Brick Company.

Heistand, Elza F., Muncie, Indiana, Superintendent, Crucible Department,
Gill Clay Pot Company.
* Died October 13, 1918.

AMERICAN CERAMIC SOCIETY. 831

lelser, Perry D., Edgewood, Maryland, Honor Roll.

lelwig, Frank H., Trinidad, Colorado, Brick and Tile Company.

lenry A. V. (Address unknown.)

lenry, Frank R., M.D., Dayton, Ohio, Manager, Dayton Grinding Wheel

Company.
lepler, I. F., Narberth, Pa., Chief Engineer, General Refractories Company.
lerbst, Abram W., Canaan, Conn., New England Slag Company.
lerron, James H., 1364 West Third St., Cleveland, Ohio, Consulting

Engineer.
less, Henry W., Toledo, Ohio, The Eibbey Glass Company,
lettinger, Edwin L., 1325 Mineral Spring Rd., Reading, Pa., Assistant

Secretary and Purchasing Agent, T. A. Willson and Company, Inc.
libbins, Thomas A., Wellsville, Ohio, General Business and Sales Manager,

The Stevenson Company,
lipp, Ralph T., Massillon, Ohio, Massillon Stone and Fire Brick Company.
lipp, William G., Massillon, Ohio, Secretary and General Manager, Massillon

Stone and Fire Brick Company,
litchins, R. E., Olive Hill, Kentucky, General Manager, General Refractories

Company,
loehn, F. J., Carlinville, Illinois, Honor Roll,
loffman, George E., Trenton, New Jersey, Secretary, Monument Pottery

Company,
[olmes, Harold W., Detroit, Michigan, Honor Roll.

[ood, B. Mifflin, Atlanta, Georgia, President, Mifflin Hood Brick Company,
[ooper, Charles N., 4157 Western Blvd., Chicago, Illinois,
[ornung, Martin R., Shenango Pottery Co., Newcastle, Pa.
[orton, J. M., Sebring, Ohio, E. H. Sebring China Company,
[ostetler, G. R., Sugar Creek, Ohio, Superintendent, Sugar Creek Clay

Products Company,
tower, H. S., Pittsburgh, Pa., Professor of Physics, Carnegie Institute of

Technology.
ludson, Charles J., 39 Kingsbury St., Worcester, Mass., Special Investigator,

Norton Company,
lull, Addis E., Jr., Crooksville, Ohio, Ceramic Engineer, A. E. Hull Pottery

Company.
lung, Shen, 52 Woodruff Ave., Columbus, Ohio,
[unt, F. Sumner, Chillicothe, Ohio, Honor Roll,
reland, Harley R., Brazil, Indiana, Assistant Secretary and Treasurer,

Hydraulic Press Brick Company,
rwin, Dewitt, East Liverpool, Ohio, Secretary, Potters' Supply Company,
ttner, Warren W., 327 N. Illinois St., Belleville, Illinois, The Malinite Com-
pany/,
ackson, H. W., DuBois. Pa., General Manager, Jackson China Company,
acobs, W. M., Charleroi, Pa., Pittsburgh, Plate Glass Company,
acquart, Charles E., South River, New Jersey, Honor Roll.

832 JOURNAL OF THE

Jeffery, Joseph A., Detroit, Michigan, President, Jeffery Dewitt Compan)
Jensen, A. Lawrence, 142 156 Green St., Brooklyn, New York, Empire China

Works

Jensen, Charles H., 142 156 Green St., Brooklyn, New York, President,
Empire China Works.

Jeppson, George N., 286 West Boylston St., Worcester, Mass., Assistant
Superintendent, Norton Company.

Johnson, A. A. V., 30 N. LaSalle St., Chicago, Illinois.

Johnston, John, 3759 W. Pine Blvd., St. Louis, Mo.

Johnston, Robert M., 707 Hamilton St., New Fairview Station, Detroit,
Michigan.

Jones, Cecil, 132 Adair Ave., Zanesville, Ohio, American Encaustic Tiling
Company.

Jones, Otis L., Ogelsby, Illinois, President and Manager, Illinois Clay Prod-
ucts Company.

Justice, Ithamar M., Dayton, Ohio, Vice-President, Manufacturers' Equip-
ment Company.

Kalbfleisch, G. C, Tiffin, Ohio, Manager, Standard Manufacturing Com-
pany.

Kanengeiser, Fred R., Poland, Ohio.

Katzenstein, Leon, Worcester, Mass., Special Investigator, Norton Com-
pany.

Kebler, Leonard, Mt. Vernon, New York, Ward-Leonard Electric Company.

Keehn, Clarence C, Canandaigua, New York, Secretary and Assistant
Manager, Lisk Manufacturing Company, Ltd.

Keeler, R. B., 2505 E. 54th St., Huntington Park, Los Angeles, California.

Keenan, J. Francis, 1007 Hoge Building, Seattle, Washington, General
Manager, Denny-Renton Clay and Coal Company.

Keese, Homer G., Litchfield, Illinois.

Kempf, John R., Detroit, Michigan, Star Corundum Wheel Company.

Kendrick, Lucius S., 513 Bearinger Bldg., Saginaw, Michigan, Vice-Presi-
dent and General Manager, Central Michigan Clay Products Company.

Kent, G. G., Detroit, Michigan, Honor Roll.

Keplinger, Robert B., Canton, Ohio, Assistant General Superintendent,
Metropolitan Paving Brick Company.

Kerr, W. B., Syracuse, New York, President, Iroquois China Company.

Keuffel, Carl W., Adams and Third Sts., Hoboken, New Jersey, Keuffel and
Esser Company.

Kier, Samuel M., 2243 Oliver Bldg., Pittsburgh, Pa., President, Kier Fire
Brick Company.

Kimble, Herman K., Vineland, New Jersey, Manager, Scientific Glass De-
partment, Kimble Glass Company.

King, B. F., Colchester, Illinois, Manager, Brick and Tile Company.

King, Earl O., Box 190, Baltimore, Maryland, American Refractories Com-
pany.

AMERICAN CERAMIC SOCIETY. 833

Kleymeyer, H. C, Office 8, Furniture Bldg., Evansville, Indiana, General
Manager, Standard Brick Manufacturing Company.

Kline, Z. C, Central Falls, Rhode Island, National Lamp Works.

Knapp, Ernest W., Y. M. C. A., Canton, Ohio.

Knight, M. A., East Akron, Ohio.

Knollman, Harry J., 28 South 34th St., Philadelphia, Pa.

Knowles, H. Homer, Box 427, East Liverpool, Ohio, Superintendent, Plants
1 and 2, Knowles, Taylor and Knowles Company.

Koch, A. L., Minneapolis, Minn., Honor Roll.

Koch, Charles F., Cincinnati, Ohio, National Sales Company.

Koch, Julius J., 3325 Carolina St., St. Louis, Mo.

Kohler, Walter J., Kohler, Wisconsin, President, J. M. Kohler Sons Com-
pany.

Kraus, Charles E., 66— 87th St., Brooklyn, New York.

Kraus, Louis P., Jr., 66 — 87th St., Brooklyn, New York, Research Engineer,
Babcock and Wilcox Company.

Krause, George, 149 So. 7th St., Zanesville, Ohio.

Krick, George M., Decatur, Indiana, General Manager, Krick Tyndall and
Company.

Krieg, Henry F., 30 North La Salle St., Chicago, Illinois, William E. Dee
Company.

Krieg, William G., 15 15 Lumber Exchange Building, Chicago, Illinois.

Kruson, I. Andrew, Victor, New York, Locke Insulator Company.

Kurtz, John C, Rochester, New York, Bausch and Lomb Optical Company.

Kurtz, Thomas N., Claysburg, Pa., Standard Refractories Company.

Lacy, Miss Mattie Lee, Denton, Texas, College of Industrial Arts.

Laird, J. S., Ann Arbor, Michigan, Assistant Professor, Chemical Labora-
tory, University of Michigan.

Lambie, J. M., Washington, Pa., Vice-President and Assistant General
Manager, Findlay Clay Pot Company.

Lamont, R. A., Jr., Salem, Ohio, General Manager, National Sanitary Com-
pany.

Landers, William F., Indianapolis, Indiana, Superintendent, V. S. Encaustic
Tile Works.

Lardin, R. H., Creighton, Pa., Assistant Chemist, Pittsburgh Plate Glass
Company.

Larkin, P. G., Denver, Colorado, Ceramic Chemist, Denver Terra Cotta
Company.

Larkins, Samuel B., Salineville, Ohio, Superintendent National China Com-
pany.

Larson, Gustaf, 1426 Carroll Ave., Los Angeles, California, Los Angeles
Press Brick Company.

Laughiin, Samuel O., Wheeling, West Virginia, President and General
Manager, Wheeling Tile Company.

Lawson, Carl H., 3 Fraternal Ave., Worcester, Mass., Norton Company.

834 JOURNAL OF THE

Lawson, Geo. G., 2525 Claybourn Ave . Chicago, III., Northwestern Terra
Cotta Company.

Lawton, Alfred V., 929 Carteret Ave , Trenton, New Jersey.

Layman, Frank E., Milwaukee, Wisconsin, Cutler-Hammer Company.

LePage, Just, Willamina, Oregon, Head Burner, Pacific Pace Brick Company.

Levings, G. V. B., Via Palisade, Nevada, Union Mines

Lewis, E. S., 29 South LaSalle St., Chicago, Illinois, Secretary-Treasurer,
Advance Terra Cotta Company.

Libman, Earl E., 912 West Nevada St., Urbana, Illinois

Lillibridge, H. D., Zanesville, Ohio, American Encaustic Tiling Company

Lin, Chi C, 22 Woodruff Ave., Columbus, Ohio.

Linbarger, S. C, Niagara Falls, New York, Carborundum Company.

Linder, Cyril S., Creighton, Pa., Pittsburgh Plate Glass Company.

Lindley, Jacob, Tiltonville, Ohio, Riverside Potteries Company.

Livingston, H. J., Buffalo, New York, National Grinding Wheel Company.

Lloyd, Harry, Irondale, Ohio, Colonial Clay and Coal Company.

Locke, F. M., Victor, New York, Manufacturer of Porcelain Insulators.

Loeffler, R. W., Frankfort, Indiana, Superintendent, Ingram-Richardson
Manufacturing Company.

Long, George E., Jersey City, New Jersey, Vice-President Joseph Dixon
Crucible Company.

Longenecker, H. L., Cambridge, Maryland, Manager, Cambridge Brick
Company.

Lord, F. G., Lewiston, Pa., Sales Manager, Pennsylvania Pulverizing Com-
pany.

Louthan, William B., East Liverpool, Ohio, Manager, Louthan Supply Com-
pany.

Lovatt, John, 1308 Brunswick Bldg., Trenton, New Jersey, Kiln Builder.

Lucas, H. J., 2525 Claybourn Ave., Chicago, Illinois, Vice-President, North-
western Terra Cotta Company.

Luke, Charles H., New Bethlehem, Pa., Climax Fire Brick Company.

Lumb, Carnote, F., 11 19 Sunset Ave., Zanesville, Ohio, American Encaustic
Tiling Company.

Luter, Clark A., Baltimore, Md., Chemist, Carr-Lowrey Glass Company.

Lyon, J. Boyd, St. Louis, Mo., Laclede-Christy Clay Products Company.

Maddison, E. A. (Address unknown.)

Maddock, A. M., Jr., Trenton, New Jersey, Vice-President, Thomas Mad-
dock's Sons Company.

Maddock, Henry E., Trenton, New Jersey, John Maddock and Sons.

Maddock, John, Trenton, New Jersey, John Maddock and Sons.

Maicas, Oscar R., 1200 Hudson St., Hoboken, New Jersey, Sales Engineer,
Didier-March Company.

Malinovszky, A., 316 Portland Ave., Belleville, Illinois, Chemical Engineer.

Malkin, William R., East Liverpool, Ohio, Salesman, B. F. Drakenfeld
& Company, Inc.

AMERICAN CERAMIC SOCIETY. 835

Malm, Arthur T., 18 Orne St., Worcester, Mass., Factory Investigator, Re-
search Laboratories, Norton Company.

Malsch, Werner, 100 William St., New York City, Ceramic Department
Manager, Roessler and Hasslacher Chemical Company.

Maltby, Alfred, Corning, New York, Superintendent, Brick Terra Cotta and
Tile Company.

Mandle, I., 1318 Wright Bldg., St. Louis, Mo., Secretary and Treasurer, The
Mandle Clay Mining Company.

Mandle, Sidney, 1318 Wright Bldg., St. Louis, Mo., Assistant Treasurer,
Mandle Clay Mining Company.

*Mandler, C. J., Toledo, Ohio, President, Allen Filter Company.

Manion, L. W., 1370 Greenfield Ave., S. W., Canton, Ohio, Designer and
Builder of Industrial Furnaces.

Manor, John M., East Liverpool, Ohio, Manager, Golding Sons Company.

Marks, Melville, 29 Broadway, New York City, Moore and Munger.

Martens, Paul, Chrome, New Jersey.

Martin, S. C, Kittanning, Pa., Kittanning Brick and Fire Clay Company.

Martz, Joseph A., Honor Roll.

Mason, F. Q., East Liverpool, Ohio, Mason Color and Chemical Company.

Mauschbaugh, Henry J., 523 North Elizabeth St., Peoria, Illinois, Carter's
Brick Yard No. 2.

Mayer, C. P., Bridgeville, Pa., C. P. Mayer Brick Company.

Mayer, Walter S., 1520 Third St., New Brighton, Pa., Assistant to the
Manager, Mayer China Company.

Maynard, T. Poole, 1321-A Hurt Bldg., Atlanta, Georgia, Consulting Geolo-
gist.

Mellor, F. G., Wheeling, West Virginia, Warwick China Company.

Meritt, Myrtle E., Pittsburgh, Pa., Instructor in Ceramics, Carnegie In-
stitute of Technology.

Metzner, Otto, Cincinnati, Ohio, Superintendent of Manufacture, Rook-
wood Pottery Company.

Middleton, Jefferson, 131 7 Rhode Island Ave., Apartment No. 304, Wash-
ington, D. C, Statistician, U. S. Geological Survey.

Miller, Donald M., 633 Monmouth St., Trenton, New Jersey, Secretary,
Crossley Machine Company.

Miller, Henry E., Chicago, Illinois, Chicago Wheel and Manufacturing
Company.

Miller, Julius J., Pittsburgh, Pa., Koppers Company, Mellon Institute.

Miller, R. W., Lt., Indianapolis, Indiana, Honor Roll.

Milligan, Frank W., Parkersburg, West Virginia, Manager, General Porcelain
Company.

Mills, H. F., Noblesville, Indiana.

Millsom, Walter C, Macomb, Illinois.

Miner, Harlan S., Gloucester City, New Jersey, Chief Chemist, Welsbach
Company.
* Died December 191 8.

836 l« >URNAL OF THE

Minor, Frederick K., Olivet Building, Pittsburgh, Pa., Sales Engineer, ( elite

Products Company

Moncrieff, James W., ('.rand Rapids. Michigan, Honor Roll
Moore, Earl J., Greenville, Ohio, Honor Roll

Moore, H. W., Perth Amboy, New Jersey, Chief Chemist, Atlantic Terra
Cotta Company.

Moorshead, T. C, Alton, Illinois, Chief Engineer, Illinois (Mass Company.

Morgan, Fred A., Honor Roll.

Morris, Paul R., 300 East 9th Ave., Tarentum, Pa., Chemist, Pittsburgh
Plate Glass Company.

Morrow, Robert P., 1513 Rockefeller Bldg., Cleveland, Ohio, Salesman,
Harbison-Walker Refractories Company.

Morton, John M., Zanesville, Ohio, Superintendent, Mosaic Tile Company.

Moses, James, Trenton, New Jersey, Mercer Pottery Company.

Mossman, P. B., 315 Union Arcade Bldg., Pittsburgh, Pa.

Motz, W. H., Akron, Ohio, Secretary and Treasurer, Colonial Sign and In-
sulator Company.

Moulding, J. M., Momence, Illinois, Tiffany Enameled Brick Company.

Moulton, D. A., 811 Riverside Ave., Wellsville, Ohio.

Muckenhirn, Charles H., 550 Chalmer Ave., Detroit, Michigan, General
Representative, Standard Sanitary Manufacturing Company, Pitts-
burgh, Pa.

Mueller, Theodore E., Louisville, Kentucky, Manager, Louisville Works,
Standard Sanitary Manufacturing Company.

Muessig, C. Nick, East Liverpool, Ohio, Salesman, B. F. Drakenfeld Com-
pany, Inc.

Mulholland, V., 41 Arch St., Hartford, Conn.

Munshaw, L. M., Terra Cotta, Illinois, Ceramist, American Terra Cotta and
Ceramic Company.

Murray, G. A., Honor Roll.

Myers, Charles H., 4981^. Poplar St., Murray, Utah, Superintendent, Utah
Fire Clay Company.

Myers, Scott P., Uhrichsville, Ohio, Superintendent, Robinson Clay Prod-
ucts Company.

McAllister, James E., Trenton, New Jersey, J. L. Mott Company.

McBean, Atholl, 31 1-3 17 Crocker Bldg., San Francisco, California, Secre-
tary, Gladding McBean and Company.

McCann, James S., Streator, Illinois, Honor Roll.

McClave, J. M., 910 American Trust Bldg., Cleveland, Ohio, General
Manager, American Fire Clay and Products Company.

McConnick, Oscar C, 7802 Carnegie Ave., Cleveland, Ohio, Chief Engineer,
Clayworkers Equipment Company.

McCoy, Nelson, Roseville, Ohio, Vice-President and General Manager,
Nelson McCoy Sanitary and Stoneware Company.

McCoy, William, 334 Adair Ave., Zanesville, Ohio, Mechanical Engineer,
American Encaustic Tiling Company.

AMERICAN CERAMIC SOCIETY. 837

IcDanel, Walter W., 532 13th Ave., New Brighton, Pa., Mould Maker,

Bureau of Standards.
IcDougal, Taine G., Flint, Michigan, Champion Ignition Company.
IcDowell, J. S., Pittsburgh, Pa., Research Department, Harbison-Walker

Company.
IcDowell, S. J., Flint, Michigan, Ceramic Engineer, Champion Ignition

Company.
IcElroy, R. H., Dayton, Ohio, International Clay Machinery Company.
IcGregor, John R., 208 S. La Salle St., Chicago, Illinois, Eagle-Picher Lead

Company.
IcHose, Malcolm M., Perth Amboy, New Jersey, Manager, L. H. McHose,

Inc.
IcKaig, W. Wallace, Cumberland, Maryland, McKaig Machinery Foundry

and Supply Works.
IcKee, Jay C, Riverside Ave., Wellsville, Ohio, Chief Clerk, McLain Fire

Brick Company.
IcKelvey, John H., St. Louis, Mo., Sales Manager, Laclede- Christy Clay

Products Company.
lacKenzie, William G., Wilmington, Delaware, Manager, Golding Sons

Company, Delaware Department.
IcKinley, J. M., Curwensville, Pa., Crescent Refractories Company.
IcLaughlin, John, Box 31, Tiltonsville, Ohio, Wheeling Sanitary Manu-
facturing Company.
lacMichael, P. S., Auburn, Washington, President, Northern Clay Com-
pany.
IcMillan, Herbert S., Detroit, Michigan, Manager and Secretary, Porcelain

Enameling and Manufacturing Company.
IcNaughton, Malcolm, Jersey City, New Jersey, Superintendent, Joseph

Dixon Crucible Company.
IcPadden, J. H., 26 Cortlandt St., New York City, Secretary Quigley

Furnace Specialties Company.
IcVay, T. N., Streator, Illinois, Streator Brick Company,
elson, Idris, 260 S. Academy St., Galesburg, Illinois, Chemist, Purington

Paving Brick Company,
iblock, Charles, Zanesville, Ohio, American Encaustic Tiling Company,
icely, Charles A., Watsontown, Pa.
ielsen, M. P., 758 Taunten Square, Zanesville, Ohio,
ies, Frederick H., D.D.S., 859 Bay Ridge Ave., Brooklyn, New York,
iles, Glenn H., Room 602, 60 Wall St., New York City, Engineer, The

Improved Equipment Company,
itchie, Charles C, Depue, Illinois, Chief Chemist, Mineral Point Zinc

Company,
oel, S. L., Bradenville, Pa., Pittsburgh High Voltage Insulator Company,
othstine, A. C, Bethesda, Maryland, Bureau of Chemistry,
ffice, Leon R., Pittsburgh, Pa., Ceramic Chemist, H. Koppers Company,

Mellon Institute.

838 JOURNAL OF THE

Olsen, Peter C, 150 Nassau St., New York City.

Olsson, L. Zach, 130 Pearl St., New York City, Consulting Chemist.

Orth, Frank, 3437 Fir Street, Indiana Harbor, Indiana, Manufacturer, Sand

Lime Brick.
Ortman, Fred B., 252.5 daybourn Ave., Chicago, Illinois, Northwestern T< rra

Cotta Co.
Osburn, Russell E., R. F. D. No. 1, Morris, Illinois, Developing Fire Clay.
Ossowski, Charles S., Cicero, Illinois, Vice-President and Superintendent,

Chicago Vitreous Fnamel Products Company.
Oudin, Charles P., 2327 Pacific Ave., Spokane, Washington, Manager,

Pacific Fire Brick Company.
Overbeck, Miss Elizabeth G., Cambridge City, Indiana, Overbeck Pottery.
Owens, Francis T., Ridgway, Pa., Treasurer and General Manager, Ridgway

Brick Company.
Owens, J. B., Zanesville, Ohio.

Parker, George W., 1624 Railway Exchange Bldg., St. Louis, Mo., Vice-
President and General Manager, Russell Engineering Company of St.

Louis.
Parkinson, J. C, 704 Third Ave., Tarentum, Pa., Chemist, Pittsburgh Plate

Glass Company.
Pass, Richard H., Syracuse, New York, Chemist, Onondaga Pottery Company.
Paterson, Alexander, Clearfield, Pa., Secretary and Treasurer, Paterson

Clay Products Company.
Pearson, Howard L., Mexico, Mo., Engineer, A. P. Green Fire Brick Com-
pany.
Pellerano, Silvio, 1837 — 7jst St., Brooklyn, New York, Assistant Chemist,

Chemical Laboratory, Hemming Manufacturing Company, Garfield,

New Jersey.
Pelton, Herbert E., 1922 Dracena Drive, Los Angeles, California, Secretary

and Treasurer, West Coast Tile Company, Inc.
Penfield, L. W., Willoughby, Ohio, Vice-President, American Clay Machinery

Company.
Penfield, R. C, 1619 Conway Bldg., Chicago, Illinois, President and General

Manager, American Clay Machinery Company.
Peregrine, Clarence R., Pittsburgh, Pa., General Superintendent, Macbeth-
Evans Glass Company.
Pettinos, George F., 305 North 15th St., Philadelphia, Pa., Senior Partner,

Pettinos Brothers, Pennsylvania Crucible Company.
Pfalzgraf, Charles F., Baltimore, Maryland, President, Baltimore Stamping

and Enameling Company.
Pfau, Charles, Cincinnati, Ohio, President, Pfau Manufacturing Company.
Phillips, William L., 1016 San Antonio Ave., Alameda, California, N. Clark

and Sons.
Pierce, O. W., Olean, New York, Olean Tile Company.
Pierce, Robert H. H., Box 516, Hazelwood, Pittsburgh, Pa., Chief Chemist,

Harbison-Walker Refractories Company.

AMERICAN CERAMIC SOCIETY. 839

Pike, Robert D., 74 New Montgomery St., San Francisco, California, Con-
sulting Chemical Kngineer.

Pitcairn, William S., 104 Fifth Ave., New York City, Importer of China and
Earthenware.

Pitcock, Lawrence, Crooksville, Ohio, Superintendent, Crooksville China
Company.

Piatt, Chauncey B., 406 Hubbell Bldg., Des Moines, Iowa.

Pohle, Louis, Trenton New Jersey, Ceramic Chemist, Monument Pottery
Company.

Pohs, Frank, Rockford, Illinois, Vice-President and Manager, Rockford
Vitreous Enameling and Manufacturing Company.

Polen, George A., 1012 Wooster Ave., Canal Dover, Ohio, General Super-
intendent, Robinson Clay Products Company.

Poole, Joshua, East Liverpool, Ohio, Manager, Homer-Laughlin China
Company.

Porter, F. B., 204>/2 Houston St., Fort Worth, Texas, President, Fort Worth
Laboratories.

Post, M. P., Commerce, Mo., Superintendent, Post Brothers.

Potter, William A., Richmond, California, Factory Manager, Pacific Porcelain
Ware Company.

Powell, William H., 200 West 70th St., New York City, President, Atlantic
Terra Cotta Company.

Preston, F. C, 509 Cuyahoga Bldg., Cleveland, Ohio, Vice-President and Sales
Manager, Dover Fire Brick Company.

Primley, Walter S., Chicago, Illinois, Honor Roll.

Pulsifer, H. M., Manhattan Bldg., Chicago, Illinois, Constructor of Tunnel
Kilns, George H. Holb and Company.

Purinton, Bernard S., Wellsville, Ohio, United States Pottery Company.

Pyatt, Frank E., 128 Sharon Ave., Zanesville, Ohio.

Quaintance, Charles F., Golden, Colorado, Secretary, Herold China and
Pottery Company.

Radabaugh, N. B., 1572 Rydal Mount Road, Cleveland, Ohio, President,
The Phoenix Hardening Equipment Company.

Rainey, E., R. F. D. 34, Peoria, Illinois, Superintendent of Yards, Carter
Brick Company.

Rainey, Lloyd, Beaver, Pa., Superintendent, Fallston Fire Clay Company.

Ramsay, Andrew, Mt. Savage, Maryland, The Andrew Ramsay Company.

Ramsay, J. D., St. Marys, Pa., President and General Manager, Elk Fire
Brick Company.

Rancke, Louis N., Lonaconing, Maryland, Consulting Engineer, Chemist.

Randall, James E., Indianapolis, Indiana, Junior Editor, "The Clayworker."

Rankin, George A., Route 1, Rosslyn, Virginia.

Rathjeins, G. W., Major, Honor Roll.

Raymond, Geo. M. (Address unknown.)

Rea, William J., Buffalo, New York, Superintendent, Buffalo Pottery Com-
pany.

s,«, JOURNAL OF THE

Reagan, F. H., Victor, New York, General Manager, Locke [nsulatoi Com

puny.
Reed, Adam, Zamsvilk', Oliio, President, Peters and Reed Pottery Compaq
Reed, Henry F., 426 Teece Aviv, Bellevue, Pa., Assistant General Manager,

Standard Sanitary Manufacturing Company.

Reeve, Howard T., [52 Main St., Southbridge, Mass, Chief Scientist, Ameri
can Optical Company.

Reynolds, Pierce B., 41,7, Wyoming Ave, Kingston, Pa

Rhead, Frederick H., Zanesville, Ohio, American Encaustic Tiling Com
pany.

Rhead, Mrs. Frederick H., Zanesville, Ohio.

Rhead, Harry W., 144 Putnam Ave., Zanesville, Ohio, Designer, Roseville
Pottery Company.

Rhoads, Ralph S., St. Louis, Mo., Evens and Howard Fire Brick Company.

Rice, Bryan A., Elyria, Ohio, Assistant Chemist, Elyria Enameled Products
Company.

Richard, L. M., 919 Yenezia Ave., Venice, California.

Richardson, Ernest, Beaver Falls, Pa., Vice-President and Treasurer, Ingram-
Richardson Manufacturing Company.

Riess, Dewitt F., Sheboygan, Wisconsin, Secretary, Yollrath Company.

Rixford, Guy L., Honor Roll.

Robertson, Fred H., 809 North Alvarado St., Los Angeles, California, Los
Angeles Pressed Brick Company.

Robertson, H. S., Pittsburgh, Pa., Honor Roll.

Robineau, S. E., 108 Pearl St., Syracuse, New York, President, Keramic
Studio Publishing Company.

Robinson, Louis G., 4th and Park Aves., Newport, Kentucky.

Rochow, W. F., Pittsburgh, Pa., Chemical Engineer, Harbison-Walker Re-
fractories Company.

Rockhold, Kenneth E., Zanesville, Ohio, Mark Manufacturing Company.

Rodgers, Eben, Alton, Illinois, Secretary and Treasurer, Alton Brick Com-
pany.

Roessler, Franz, 39 High St., Perth Amboy, New Jersey, Roessler and Hass-
lacher Chemical Company.

Rogers, Frederick W., Beaver Dam, Wisconsin, Malleable Iron Range Com-
pany.

Ross, Donald W., Washington, Pa., Findlay Clay Pot Company.

Rusoff, Samuel, Anderson, Indiana, National Tile Company.

Russel, John W., 522 Chestnut St., Columbia, Pa., President, Marietta
Hollow Ware and Enameling Company.

Ryan, John J., McQueeny, Texas, Seguin Brick and Tile Company.

Salisbury, Bert E., 18 10 W. Genesee St., Syracuse, New York, President,
Onondaga Pottery Company.

Sanders, John W., Moundsville, West Virginia, Enameler, United States
Stamping Company.

Si

AMERICAN CERAMIC SOCIETY. 841

5ant, Richard C, East Liverpool, Ohio, Secretary, The John Sant and Sons
Company.

Sant, Thomas H., East Liverpool, Ohio, Secretary-Treasurer, The John Sant
and Sons Company.

Sassetti, F. L., 4147 Van Buren St., Chicago, Illinois.

Saunders, William E., 1401 Arch St., Philadelphia, Pa., Engineer, United
Gas Improvement Company.

Savage, Harold R., Worcester, Mass., Refractories Chemist, Norton Com-
pany.

Schaeffer, John A., Joplin, Mo., General Superintendent, Eagle-Pieher Lead
Company.

Schermerhorn, Joseph A., 938 Carteret Ave., Trenton, New Jersey, Trenton
Porcelain Company.

Schaulin, George M., Streator, Illinois, Honor Roll.

Schmid, A. F., Zanesville, Ohio, Superintendent, Zanesville Art Pottery Com-
pany.

Schmidt, Henry, 100 William St., New York City, Salesman, Roessler and
Hasslacher Chemical Company.

Scholes, Samuel R., Ph.D., Rochester, New York, Chemist, H. C. Fry Glass
Company.

Schory, Virgil S., 231 Hudson St., Tiffin, Ohio, Ceramic Engineer, Standard
Sanitary Manufacturing Company.

Schulze, John F. W., 43 Ridge Drive, Yonkers, New York.

Schwetye, Fred H., St. Louis, Mo., Superintendent, Laclede-Christy Clay
Products Company .

Scott, M. R., 102 Virginia Ave., Rochester, New York, Bausch and Lomb
Optical Company.

Scott, W. C, Warrior, Alabama, Superintendent, Sibley Brick Company.

Seasholtz, J. M., Front and Spruce Sts., Reading, Pa., Owner, Porcelain
Enameling Plant.

Seaver, Kenneth, Pittsburgh, Pa., Harbison-Walker Refractories Company.

Sebring, Charles L., Sebring, Ohio, General Manager, Sebring, Pottery.

Jewell, Sidney I., Pittsburgh, Pa., Laboratory Assistant, Bureau of Stand-
ards.

sharp, Donald E., Hamburg, New York, Spencer Lens Company.

Sheehy, J. F., Newport, Kentucky, Superintendent and Manager, Alhambra
Tile Company.

Sheerer, Mary G., New Orleans, La., Assistant Director, Newcomb Pottery.

Sherban, Daniel V., 1 103 West 2nd St., Oil City, Pa.

Shoemaker, George W., Perrysville, Ohio,

Silver, Miss Anna K., 9 Hawthorne St., Worcester, Mass., Analytical Division,
Norton Company.

Sinclair, Herbert, Trenton, New Jersey, General Manager, Star Porcelain
Company.

Singer, L. P., Lincoln, California, Chemist, Gladding McBean and Company.

842 JOURNAL OF THE

Sloan, Alex, Box 410, Cumberland, Maryland, Manufacturing Chemist.

Smith, A. M., 401 Columbia Bldg., Portland, Oregon.

Smith, Harry W., Box ,U><>, Cleveland, Ohio, Roessler and Hasslai bi 1 Chemical

Company.

Smith, James M., New Castle, Pa., Treasurer, ShenangO Pottery Company

Smith, Maurice A., Jeanette, Pa., Vice Presided and General Manager,
McKee Class Company.

Smith, Norman G., Brunswick, Me., Manager and Treasurer, Maim- Feldspar
Company.

Smith, Perry A., New Brighton, Pa., Secretary, A. F. Smith Company.

Smith, Will L., Jr., Chester, West Virginia, General Manager, Taylor Smith
and Taylor Company.

Solon, A. L., 195 Fremont St., San Jose, California.

Solon, Marc, Trenton, New Jersey, General Manager, Mercer Pottery Com-
pany.

Sortwell, Harold H., 309 East John St., Champaign, Illinois.

Speir, Harry F., Cedarville, New Jersey, President, New Jersey Pulverizing
Company.

Springe, Otto, St. Louis, Mo., Honor Roll.

Sproat, Ira E., Sebring, Ohio, Honor Roll.

Stallings, A. G. T., Mexico, Mo., Superintendent, A. P. Green Fire Brick
Company.

Stanger, Frederick, Real Estate Trust Bldg., Philadelphia, Pa., Sales
Manager, Enterprise White Clay Company.

Stangl, J. M., Dundee, Illinois, Superintendent, Haeger Potteries.

Stanley, Wm. W., New Brunswick, New Jersey, Wasson Piston Ring Com-
pany.

Stapler, Walter S., Stevens Pottery, Georgia, General Manager, Stevens
Brothers and Company.

Staudt, August, Perth Amboy, New Jersey, President, Perth Amboy Tile
Works.

Steinhoff, F. L., Chicago, Illinois, Honor Roll.

Stepan, Alfred C, Room 155 1, Conway Bldg., Chicago, Illinois, Manager,
Roessler and Hasslacher Chemical Company.

Stern, Newton W., 67 New Montgomery St., San Francisco, California,
Secretary and Treasurer, Pacific Porcelain Ware Company.

Stevens, Douglas F., Cayuga, Indiana, Superintendent, Acme Brick Com-
pany.

Stevenson, Wm. G., Oregon, Illinois, Secretary-Treasurer, Ohio Silica Com-
pany.

Stewart, Andrew H., Pittsburgh, Pa., Mellon Institute.

Stockton-Abbott, Lyle, Hotel Plaza, Port Arthur, Texas, Engineer.

Stone, Charles A., Box 320, Kenova, West Virginia, Basic Products Company.

Stone, George C, 55 Wall St., New York City, New Jersey Zinc Company.

Stoneman, William N., Charleroi, Pa., Chemist, Macbeth-Evans Glass Com-
pany.

AMERICAN CERAMIC SOCIETY. 843

Stowe, Charles B., Cleveland, Ohio, President, National Fire Brick Com-
pany.
Sullivan, Eugene C, Ph.D., Corning, New York, Chief Chemist, Corning

Glass Works.
Swalm, Phaon H., Trenton, New Jersey, 344 Bellevue Ave.
Sweely, B. T., Chicago, Illinois, Western Electric Company, Research

Laboratories.
Swift, George C, 2029 E. 115th St., Cleveland, Ohio, Superintendent, Enamel

Products Company.
Tamlyn, Walter I., 359 Crown St., Brooklyn, New York, Assistant Engineer

with Ralph D. Mershon.
Tatton, Joseph, Y. M. C. A. Bldg., South Amboy, New Jersey.
Taylor, Royal W., 707 — 12th St., N. W., Canton, Ohio, Assistant Secretary

and Chemist, Canton Stamping and Enameling Company.
Tefft, C. Forrest, Darlington, Pa., Manager, Darlington Clay Products Com-
pany.
Tefft, T. Dwight, Zanesville, Ohio, Ceramist, Mosaic Tile Company.
Thomas, Chauncey R., 2336 San Pablo Ave., Berkeley, California, 'The Tile

Shop."
Thomas, George E., St. Louis, Mo., General Superintendent, Highlands Fire

Clay Company.
Thomas, George W., East Liverpool, Ohio, President, R. Thomas and Sons

Company.
Thomas, James R., Crawfordsville, Indiana, Manager, Standard Brick Com-
pany.
Thompson, Dale, East Liverpool, Ohio, Treasurer, C. C. Thompson Pottery

Company.
Thompson, Harry M., Washington, Pa., Furnace Engineer, Hazel-Atlas

Glass Company.
Thurlimann, Leo, 924 Fullerton Ave., Chicago, Illinois.
Thwing, C. B., Ph.D., Philadelphia, Pa., President, Thwing Instrument

Company.
Tillotson, George S., 43 Water St., Tiffin, Ohio, Sterling Grinding Wheel

Company.
Tilton, C. B., Worcester, Mass., Honor Roll.
Tilton, Earl, 1555 Belmont Ave., Columbus, Ohio, Columbus Forge and Iron

Company.
Timmerman, Walter F., Kansas City, Kansas, Vice-President, Western Terra

Cotta Company.
Tone, Frank J., Niagara Falls, New York, Works Manager, Carborundum

Company.
Townsend, Everett, 344 Bellevue Ave., Trenton, New Jersey, General

Manager, Robertson Art Tile Company.
Treischel, Chester, Schenectady, New York, General Electric Company.
Trifonoff, Boris, Zanesville, Ohio, American Encaustic Tiling Company.

si i l' fl RNAL I »!•' THE

Trowbridge, Prentiss S., [337 Kings Highway, St. Louis, Mo., Hydraulic

Pressed Brick Company
Truby, H. A., Creighton, Pa . Research Department, Pittsburgh Plate i

company.
Truman, Gail R., 5801 Manchester Ave., St. Louis, Mo., Ceramic Chemist,

St Louis Terra Cotta Company.
Tucker, Gus M., Akron. Ohio, Chemist, Robinson Clay Products Company,

Factory No. 1.
Tiirk, Karl, 126 S. Patterson Park Ave., Baltimore, Maryland. Porcelain

Enamel and Manufacturing Company.
Turner, A. M., Sciotovillc, Ohio, Scioto Fire Brick Company.
Turner, James, 1004 Stuyvesant Ave.. Trenton, New Jersey, Manager, Cook

Pottery Company.
Unger, J. S., 1054 Prick Annex, Pittsburgh, Pa.

Vance, George E., Springfield, Ohio, Safety Emery Wheel Company.
Vane, A. S., Bridesburgh, Philadelphia, Pa., The Abrasive Company.
Van Schoick, E. H., Honor Roll.
Vodrey, William E., East Liverpool, Ohio, General Manager, Yodrey Pottery

Company.
Vogel, C. J., Irondale, Ohio, The McLain Fire Brick Company.
Vogt, C. C, Pittsburgh, Pa., Mellon Institute.
Vollkommer, Josef, Bessemer Bldg., Pittsburgh, Pa., Manager and President,

Vitro Manufacturing Company.
Vollmer, August, Jr., St. Louis, Mo., St. Louis Pottery and Manufacturing

Company.
Volz, Wm. J., .St. Louis, Mo., Evens and Howard Fire Brick Company.
Wagner, C. L., Seattle, Wash,. Superior Portland Cement Company.
Wagner, Fritz, 1049 Oakdale Ave., Chicago, Illinois, Vice-President and

Treasurer, Northwestern Terra Cotta Company.
Waite, V. R., 225 Hurt Bldg., Atlanta, Georgia.
Walcott, A. J., 41 Prince St., Rochester, New York, Research Physicist,

Bausch and Lomb Optical Company.
Walden, Albert S., Cleveland, Ohio, Foreman, National Carbon Company.
*Walduck, Charles L., Worcester, Mass., Norton Company.
Walker, Charles H., 400 West 4th St., East Liverpool, Ohio, Assistant Super-
intendent, Homer Laughlin China Company.
Walker, Prescott H., Niagara Falls, New York, Honor Roll.
Walton, H. K., 59 Pearl St., New York City, New York Agent, Thwing

Instrument Company.
Walton, Samuel F., Niagara Falls, New York, Assistant Ceramic Depart-
ment, Carborundum Company.
Weaning, A. Marlowe, 1360 Nicholson Ave., Cleveland, Ohio, Cowan Pottery.
Weber, August, Jr., 1 Stratford Road, Schenectady, New York, President,

Weber Electric Company.

* Died October 14, 1918.

i

AMERICAN CERAMIC SOCIETY. 845

Wegener, Harvey A., East Pittsburgh, Pa., Commercial Engineer, Westing-
house Electric Company.

Weigel, Chas., Hebron, N. Dak., Hebron Fire and Pressed Brick Company.

Weil, Edgar H., Cleveland, Ohio, Vitreous Enameling Company.

Wells, Harry B., Canton, Ohio, Superintendent, Belden Brick Company.

Whelden, Frank H., Detroit, Michigan, Detroit Grinding Wheel Company.

Whitaker, Fred A., Keasbey, New Jersey, Superintendent, General Ceramics
Company.

White, Ray H., Niagara Falls, New York, Research Engineer, Norton Com-
pany.

Whitehead, Fred, 747 New York Ave., Trenton, New Jersey, Superintendent,
Electrical Porcelain and Manufacturing Company.

Whitehead, Ralph R., Woodstock, New York, Byrdcliff Pottery.

Whitelaw, James C, 1220 Boren Ave., Seattle, Washington.

Whitmer, J. D., 9 Ball St., Zanesville, Ohio, American Encaustic Tiling'
Company.

Wigfield, C. L., 1415 Middle Ave., Elyria, Ohio, Elvria Enameled Products
Company.

Wilder, T. M., Ravinia, Illinois, Proprietor, Wildwood Shop.

Wilkinson, George D., Chicago, Illinois, Vice-President, Cribben and Sexton
Company.

Wilkinson, Samuel, Kenova, West Virginia, Assistant Superintendent,
Sanitary Manufacturing Company.

Will, Otto W., Perth Amboy, New Jersey, Superintendent, Color Depart-
ment, Roessler and Hasslacher Chemical Company.

Willetts, H. G., South 10th St., Pittsburgh, Pa., The Willetts Company.

Williams, John A., Prospect St. and P. & R. R. R., Trenton, New Jersey,
Factory Manager, Mitchell-Bissell Company.

Wilson, Hewitt, Columbus, Ohio, Professor in Ceramic Engineering De-
partment, Ohio State University.

Wood, A. T., Kenova, West Virginia, General Manager, Basic Products
Company.

Woods, William J., Lewistown, Pa., Assistant General Manager, Pennsyl-
vania Pulverizing Company.

Worsham, Herman, 1144 Prudential Bldg., Buffalo, New York, Sales Engi-
neer, Carrier Engineering Corporation.

Worth, S. Harry, 404 Franklin Trust Bldg., Philadelphia, Pa., President,
Pennsylvania Feldspar Company.

Wright, George, St. Louis, Mo., Factory Superintendent, Buck Stove and
Range Company.

Wright, Joseph W., Pittsburgh, Pa., Bureau of Standards.

*Yates, Alfred, Shawmut, Pa., General Manager, Shawmut Paving Brick
Company.

Yearsley, Howard L., Haddonfield, New Jersey.
* Died April 17, 19 18.

846 JOURNAL OF THE

Yoshioka, Tosaku, Ph.D., 520 West 122nd St., N't w York City.

Young, C. B., Newark, Ohio, Secretary-Treasurer and General Manager,
< )liio State Brick and Stone Company.

Young, George F., Zanesville, Ohio, Manager, Roseville Pottery Company.

Young, Russell T., Zanesville, Ohio, Secretary, Roseville Pottery Company.

Yowell, J. B., Dudley, Illinois, Honor Roll.

Zakharoff, Alexis, 514 W. 114th St., New York City.

Zeiller, Oscar F., New York City, Secretary, B. F. Drakenfeld and Com-
pany, Inc.

Zimmerli, William F., Ph.D., Rochester, New York, Chemist, The Pfaudler
Company.

Zopfi, Albert S., Toledo, Ohio, Secretary, Buckeye Clay Pot Company.

Zwermann, Carl H., 22 Wyoming St., Newark, Ohio.

Zwermann, Theodore, 113 Linden Ave., Newark, Ohio, Enameling Super-
intendent.

FOREIGN ASSOCIATE MEMBERS

Anderson, Olaf, Statsgeolog, Mineralogisk Museum, Kristiania, Norway,
Norwegian Government Geologist, Director of Research Laboratory.

Barrett, Maurice, 17 Gledhow Ave., Leeds, England.

Bigot, A., Ph.D., 112 Ave. de Suffern, Paris, France, Refractories.

Boswell, P. G. H., Ph.D., The University, Liverpool, England, Refractories.

Broderick, J. C, Montreal, Quebec, Canada, Canadian China Clay Company.

Buckner, O, S., 787 Kaniyamura, Hiroshima, Japan, Norton Company.

Carter, Owen, Wykeham, Poole, England, Carter Encaustic Tile Works.

Cleverly, Wm. B., Jr., Avonmouth, near Bristol, England, cr H. M. Factory.

Cobb, John W., Leeds, Yorkshire, England, Professor, Fuel Department,
The University.

Coulter, Allen S., Reisholz bei Dusseldorf, Germany, cr Deutsche-Carborun-
dum Werke.

Cox, Paul E., 3 Avenue Alphand, Paris, France, cr Compagnie Generale des
Meules Maxf Grinding Corporation, Chester, Mass.

Davidson, T. R., Box 700, Montreal, Canada, Thomas Davidson Manu-
facturing Company.

Deb, S., 45 Taugra Road, Calcutta, India, Manager, Calcutta Pottery Works.

DeLuze, Henri, Avenue de Poitiers, Limoges, France, Haviland Porcelain
Company.

Dingledine, H. F., Aldershot, Ontario, Canada, Factory Manager, National
Fire Proofing Company of Canada, Ltd.

Emery, George, 4th Ave., West, Hamilton, Ontario, Canada, Canadian
Porcelain Company.

Foley, Fenwick D., Loch Lomond Road, St. John, N. B.

Fredriksson, Nils, Svedala, Sweden, Ceramic School.

AMERICAN CERAMIC SOCIETY. 847

Fujioka, Koji, Kyoto, Japan, Shofu Porcelain Manufacturing Company.

Gaby, F. A., 190 University Ave., Toronto, Canada, Chief Engineer, Hydro-
Electric Power Company.

Goddard, W. D., Hamilton, Canada, Canadian Porcelain Company.

Griffin, Carl H., Wesseling bei Koeln, Germany, Deutsche Norton Gessel-
shaft.

Groocock, Miss Alice, 865 Bathurst St., Toronto, Canada, Toronto Technical
School, Pottery Teacher.

Harvey, George R., 448 Barton St., Hamilton, Ontario, Canada, Vice-Presi-
dent and Manager, Canadian Hart Wheels, Ltd.

Haviland, Jean, Limoges, France, Haviland & Company.

Hayhurst, Walter, Woodland, Accrington, England.

Hirano, Kosuke, Tokyo, Japan, Tokyo-Koto-Koggo-Grakko.

Hodson, G. A., 58 Park Road, Loughborough, England, Managing Director,
Hathern Station Brick and Terra Cotta Company, Ltd.

Horsley, Thomas N., Honor Roll.

Hoursouripe, J., Palantelen — F. C. C. G. B. A., Republica Argentina.

Hurlburt, Frederick H., 72 Howland Ave., Toronto, Ontario, Canada, Roches-
ter Feldspar Mills.

Ide, Kiyoshi, Amagasaki City, Japan, The Amagaski Factory, Asahi Glass
Company.

Kato, Mitsu, Uno-ko, Okayamaken, Japan, Managing Director, Uno Fire
Brick Company.

Kemp, W. A., Toronto, Canada, Vice-President, Sheet Metal Products
Company.

King, C. A., Near Leeds, Yorkshire, England, Farnley Iron Company, Ltd.

Kitamura, Y., Kyoto, Japan, Director and Chief Engineer, Shofu Industrial
Company, Ltd.

Kondo, S., Tokio, Japan, Tokio Higher Technical School.

Kurahashi, T., P. O. Terasho, Shigaken, Japan.

Lauer, Frank E., 2389 Mance St., Montreal, Canada.

Love, Herbert G., 207 Hammond Bldg., Moose Jaw, Canada, Dominion
Fire Brick and Clay Products Company, Ltd.

Lundgren, H. J., Peterboro, Ontario, Canada, Superintendent, Canada Gen-
eral Electric Company.

Lusby, Charles A., Amherst, Nova Scotia, Amherst Foundry Company.

Marson, Percival, Norton Park, Edinburgh, Scotland, Edinburgh and Leith
Glass Works.

Mellor, J. W., Ph.D., Sandon House, Regent St., Stoke-on-Trent, England,
Secretary, English Ceramic Society.

Millar, John B., Clayburn, B.C.

Misumi, Aizo, Tobata, near Moji, Japan, Asahi Glass Company.

Momoki, Saburo, 84 Kobayashi-Cho, Nagoya, Japan, Ceramic Engineer.

McMaster, Arthur W., Box 736, Sidney, Nova Scotia, Dominion Iron and
Steel Company.

848 AMERICAN CERAMIC SOCIETY

Okura, K., S4 ECobayashi Cho, Nagoya, Japan, Directing Manager of Japan

l'oi ((.lain Corporation.

Paulsen, Carl A., 45 Smallegade, Copenhagen F, Denmark.

Pike, Leonard G., Wareham, Dorsetshire, England, Pikes C,lay Mines

Ramsden, C. E., Honor Roll.

Rogers, Gregory L., Sault vSte. Marie, Canada, Ceramic Engineer, Algoma

Steel Corporation.
Sailly, Paul, 64 Rue Franklin, Ivry-Port, Seine, Prance, Compagnie General

D'Klectricite.
Saxe, Charles W., Gatun, C. Z., Norton Company.
Segsworth, W. E., 103 Bay St., Toronto, Canada, Consulting Engineer,

Pennsylvania Feldspar Company.
Shanks, Douglas, Barrhead, near Glasgow, Scotland, Shanks and Company,

Ltd., Victorian Pottery.
Shanks, John, Barrhead, near Glasgow, Scotland, Shanks and Company,

Ltd., Victorian Pottery.
Shanks, John A. G., Glasgow, Scotland, Killed in action, Oct. 3, 1917. Honor

Roll.
Spier, Charles W., London, S. W., England, Director, The Morgan Crucible

Company, Battersea Works.
Sweet, George, The Close, Wilson St., Brunswick, Victoria, Australia, Pro-
prietor, Brunswick Brick, Tile and Pottery Works.
Thomas, Charles W., Clifton House, Old Swinford, Stourbridge, Kngland,

Honor Roll.
Tooth, W. E., Woodville, Burton-on-Trent, England, Director, Bretby Art

Pottery.
Turner, W. E. S., D.Sc, Sheffield, England, Director Department Glass

Technology, The University.
Umeda, Otogoro, Tokio, Japan, Shinagawa Hakurengwa-Kwaisha Shinagawa.
Villalta, J. F. R., Barcelona, Spain, American Clay Machinery Company.
Walker, E. E., Mitcham, Victoria, Australian Tesselated Tile Company.
Ward, S. Paul, Rancagua, Chile, S. A., Braden Copper Company.
Wyse, Henry T., 106 Braid Road, Edinburgh, Scotland, Art Master, Ladies'

College.
Yamada, Sanjiro, Tokio, Japan, Factory Manager, Asahi Glass Company.
Yamamoto, Tamesburo, 2, Yorikicho, Osaka, Japan, President, Yamatame

Glass Manufacturing Company, Osaka Yamasan and Company, Kobe.
Yuill, John W., 422 1st St., S. E., Medicine Hat, Alberta, Canada, Chemist.

"Hn flfcemoriam"

Shanks, John A. G. (Killed in Action)

Yates, Alfred Walduck, Charles L.

Mandler, C. J.

Hansen, Harry (Died in Service)

850 JOURNAL OF THE

CONTRIBUTING MEMBERS.

Abrasive Company, Tacony and Fraley Sts., Bridesburg, Philadelphia, Pa.

American Dressier lunnel Kilns, Inc., 171 Madison Ave., New York City.

American Emery Wheel Works, Providence, Rhode Island.

American Encaustic Tiling Company, Zanesville, Ohio.

The American Porcelain Company, Bast Liverpool, Ohio.

American Terra Cotta and Ceramic Company, People's Gas Bldg., Chicago,

Illinois.
American Trona Corporation, 724 South Spring St., Los Angeles, California.
The Ashland Fire Brick Company, Ashland, Kentucky.
Bausch and Lomb Optical Company, Rochester, New York.
Benjamin Electric Manufacturing Company, Chicago, Illinois.
Brick and Clay Record, 610 Federal St., Chicago, Illinois.
Buckeye Clay Pot Company, Toledo, Ohio.
The Canton Stamping and Enameling Company, Canton, Ohio.
Champion Ignition Company, Flint, Michigan.
The Colonial Company, Last Liverpool, Ohio.
Cortland Grinding Wheel Company, Cortland, New York.
The Davidson-Stevenson Porcelain Company, East Liverpool, Ohio.
B. F. Drakenfeld and Company, Inc., 50 Murray St., New York City.
Dunn Wire Cut Lug Brick Company, Conneaut, Ohio.
East Liverpool Potteries Company, Wellsville, Ohio.
The Edgar Plastic Kaolin Company, Metuchen, New Jersey.
Elyria Enameled Products Company, Elyria, Ohio.
Findlay Clay Pot Company, Washington, Pa.
The French China Company, Sebring, Ohio.
Frink Pyrometer Company, Lancaster, Ohio.
The Golding Sons Company, East Liverpool, Ohio.
The Hall China Company, East Liverpool, Ohio.
Hanovia Chemical and Manufacturing Company, Chestnut St., and N. J.

R. R. Ave., Newark, New Jersey.
Harbison- Walker Refractories Company, Pittsburgh, Pa.
The Harker Pottery Company, East Liverpool, Ohio.
The Harshaw, Fuller and Goodwin Company, Cleveland, Ohio.
Homer-Laughlin China Company, East Liverpool, Ohio.
Iroquois China Company, Syracuse, New York.
Jeffery-Dewitt Company, Detroit, Michigan. •
Edwin M. Knowles China Company, Newell, West Virginia.
Knowles, Taylor and Knowles Company, East Liverpool, Ohio.
The Limoges China Company, Sebring, Ohio.
Louthan Supply Company, East Liverpool, Ohio.
Maine Feldspar Company, Auburn, Maine.

AMERICAN CERAMIC SOCIETY. 851

F. Q. Mason Color and Chemical Company, East Liverpool, Ohio.

Massillon Stone and Fire Brick Company, Massillon, Ohio.

Maxf Grinding Wheel Corporation, Chester, Mass.

The Mosaic Tile Company, Zanesville, Ohio.

D. E. McNicol Pottery Company, East Liverpool, Ohio.

T. A. McNicol Pottery Company, East Liverpool, Ohio.

Norton Company, Worcester, Mass.

Ohio Pottery Company, Zanesville.Ohio.

The Onondaga Pottery Company, Syracuse, New York.

Owen China Company, Minerva, Ohio.

Pennsylvania Salt Manufacturing Company, Pittsburgh, Pa.

Perth Amboy Tile Company, Perth Amboy, New Jersey.

The Pfaudler Company, Rochester, New York.

Pittsburgh High Voltage Insulator Company, Derry, Pa.

The Pittsburgh Plate Glass Company, Pittsburgh, Pa.

Potters Supply Company, East Liverpool, Ohio.

Roessler and Hasslacher Chemical Company, 100 William St., New York City.

John H. Sant and Sons Company, East Liverpool, Ohio.

The Saxon China Company, Se bring, Ohio.

The Sebring Pottery Company, Sebring, Ohio.

Smith-Phillips China Company, East Liverpool, Ohio.

Standard Pottery Company, East Liverpool, Ohio.

Star Porcelain Company, Trenton, New Jersey.

Steubenville Pottery Company, Steubenville, Ohio.

Streator Clay Manufacturing Company, Streator, Illinois.

The Taylor, Smith and Taylor Company, Chester, West Virginia.

R. Thomas and Sons Company, East Liverpool, Ohio.

C. C. Thompson Pottery Company, East Liverpool, Ohio.

The Trenle China Company, East Liverpool, Ohio.

Veritas Firing System, Prospect Laboratories, Trenton, New Jersey.

Vitro Manufacturing Company, Bessemer Bldg., Pittsburgh, Pa.

Vodrey Pottery Company, East Liverpool, Ohio.

Warwick China Company, Wheeling, West Virginia.

The West End Pottery Company, East Liverpool, Ohio.

JOURNAL OF THE

MEMBERSHIP.

The preceding lists have been brought up-to-date, January i, 1919. Owing
to the well organized and executed campaign on the part of the Committei
on Membership for new members, the- enrollment of the Society lias under-
gone an unusual and highly gratifying growth. This has been 111 spite of the
disturbed conditions prevailing abroad and augurs well for the future welfare
and prosperity of the Society.

The following analysis is interesting:
Membership, February i, 191 7.

Honorary members 4

Active members, resident 71

Active members, foreign 5

Associate members, resident 414

Associate members, foreign 43

Total 537

Membership, February i, 191 8.

Honorary members 4

Life members 1

Active members, resident 79

Active members, foreign 9

Associate members, resident 568

Associate members, foreign 56

Contributing members 33

Total . 750

Membership, January i, 1919.

Honorary members 3

Life members 1

Active members, resident 99

Active members, foreign 6

Associate members, resident 722

Associate members, foreign 72

Contributing members 73

Total 976

AMERICAN CERAMIC SOCIETY.

853

HONOR ROLL MEMBERS.

Allen, F. B.
Anderson, Robert E.
Arnold, Howard C.
Ayer, Kenneth R.
Bacon, R. F.
Balmert, Richard M.
Bartells, H. H.
Bates, Charles E.
Best, Harold A.
Brockbank, Clarence J.
Broga, Wilson C.
Brown, Wilbur F.
Bucher, L. R.
Byers, Louis L.
Carter, Benjamin F.
Chase, John A.
Chormann, O. I.
Clayter, Frederic C.
Cowan, R. Guy
Crew, H. F.
Dorst, Max
Ferguson, Richard D.
Ferguson, Robert F.
Fisher, G. P.
Fulton, Kenneth I.
Geiger, Chas. F.
Gerber, Albert C.
*Hansen, Harry
Helser, Perry D.
Hoehn, F. J.
Holmes, Harold W.
Hope, Herford
Horning, Roy A.

* Died in service.

Hornung, M. R.
Horsley, Thomas N.
Howat, Walter L.
Hunt, F. Sumner
Jacquart, Charles E.
Kent, G. G.
Koch, A. L.
Martz, Joseph A.
Miller, R. L.
Moncrieff, James W.
Moore, Earl J.
Moore, Joseph K.
Morgan, Fred A.
Murray, G. A.
McCann, James S.
Orton, Edward, Jr.
Primley, Walter S.
Ramsden, C. E.
Rathjeins, G. W.
Rixford, Guy L.
Robertson, H. S.
Schaulin, George M.
Springe, Otto
Sproat, Ira E.
Steinhoff, F. L.
Thomas, Charles W.
Tilton, C. B.
Van Schoick, E. H.
Walker, Francis W., Jr.
Walker, Prescott H.
Whitford, Wm. G.
Yowell, J. B.

854

AMERICAN CERAMIC SOCIETY

HONOR ROLL STUDENT BRANCHES.

New York State Student Branch.

Ayars, Lister
Blumenthal, George
Crawford, George
Hagar, Donald
Harrington, Henry W.
Kane, Donald
Kenyon, Spicer

King, Walter
Lobaugh, Frank E.
Negus, Wayland
Reid, W. Harold
Sheppard, Mark
Sherwood, Robert F.

Ohio State University Student Branch.

Blum, J. W.
Callahan, H. D.
Cramer, W. E.
Davis, H. E.
Dodd, C. M.
Foster, H. D.
Gregorius, J. S.
Gunther, F. W.
Hall, Rex
Hartford, F. M.
Jones, R. E.
Kirschner, C. F.
Koos, E. K.
Lintz, E. H.

Mauck, F. J.
Merritt, L. M.
Minor, J. A.
Minton, G. Z.
Ramsay, John
Robinson, H. A.
Shaw, A. E.
Sieverling, P. A.
Skinner, E.
Stephens, W. H.
Stowe, G. T.
Vance, E. D.
Watkins, R. T.
Zwerner, C. G.

University of Dlinois Student Branch.

Baker, Earl B.
Bates, Charles E.
Byers, Louis L.
Carter, Benjamin F.
Chittenden, Robert M.
Christopher, Arthur B.
Davis, Raymond E.
Golden, Dios E.
Greenbaum, Charles S.

Hummeland, Ralph W.
Kennelley, Griffith S.
Kling, Carl L.
Leibson, Jacob S.
Moncrieff, James W.
Sortwell, Harold H.
Twells, Robert
Zalaski, John T.

RULES OF THE SOCIETY.

[Revised 191 8.]

I.
OBJECTS.

The objects of the American Ceramic Society are to develop the ceramic
arts and the sciences related to the silicate industries by means of meetings
for the reading and discussion of papers, the publication of scientific litera-
ture, and other activities.

II.

MEMBERSHIP.

The Society shall consist of Honorary Members, Active Members, Asso-
ciate Members, Affiliated Members, and Contributing Members.

Honorary Members must be persons of acknowledged professional emi-
nence whom the Society wishes to honor in recognition of their achievements
in ceramic science or art. Their number shall at no time exceed two per
cent of the combined active and associate membership.

Active Members must be persons competent to fill responsible positions in
ceramics. Only Associate Members shall be eligible to election as Active
Members and such election shall occur only in recognition of attainments
or of services in the science, technology, or art, of the silicate industries.

Associate Members must be persons interested in the silicate or allied in-
dustries.

Contributing Members must be persons, firms, or corporations, who, be-
ing interested in the Society, make such financial contributions for its sup-
port as are prescribed in Section III.

Honorary Members shall be nominated for election by at least five Active
Members and approved by the Board of Trustees. Their nomination shall
be placed before the Society at an annual meeting and to be elected they must
receive the affirmative vote of at least ninety per cent of those voting, by
letter ballot, at the next succeeding annual meeting.

To be promoted to active membership Associate Members must be nomi-
nated in writing by an Active Member, seconded by not less than two other Ac-
tive Members, and approved by the Board of Trustees. Their nomination
must be accompanied by a statement of their qualifications and to be elected
they must receive an affirmative vote of the majority of the members of the
Board of Trustees.

A candidate for associate membership must make application upon a form
provided by the Board of Trustees, which shall contain a written statement

JOURNAL OF THE

ol the age and professional experience <>f the candidate and a pledge to con-
t< it in to tin- laws, rules, and requirements of the Society. Such application
must be endorsed by two Active Members <>f the Society as sponsors and must
be approved by the Hoard of Trustees, The Hoard may act by lettei ballot
upon such application at any time, after which and upon payment of the fees

and dues prescribed in Section III, an approved candidate may be enrolled

on the appropriate list of the .Society.

Affiliated Members are those members of a Local Section who are not
otherwise connected with the Society. Their rights and privileges relate
solely to the affairs of the Local Sections to which they belong.

Contributing Members may be enrolled on the appropriate list of the So-
ciety at any time upon the payment of the dues prescribed in Section III.

All Honorary Members, Active Members, Associate Members, and Contrib-
uting Members shall be equally entitled to the privileges of membership,
except that only Active Members shall be entitled to vote and hold office.
The roster of each of these grades of membership shall be printed separately
in at least one publication issued by the Society annually. Any person may
be expelled from any grade of the membership of the Society if charges signed
by five or more Active Members be filed against him and if a majority of the
Board of Trustees examine into said charges and sustain them. Such per-
son, however, shall first be notified of the charges against him and be given
a reasonable time to appear before the Board of Trustees or to present a
written defense, before final action is taken.

m.

DUES.

Honorary Members shall be exempt from all fees or dues.

The initiation fee of Active Members shall be ten dollars and if this be not
paid within three months after the date of election, the latter shall be null
and void. The annual dues shall be fixed by the Board of Trustees and shall
not exceed ten dollars, four dollars of which shall be a subscription to the
Journal of the American Ceramic Society.

The initiation fee of Associate Members shall be five dollars, payable
within three months after date of election. The annual dues shall be fixed
by the Board of Trustees but shall not exceed five dollars, four dollars of which
shall be a subscription to the Journal of the American Ceramic Society.

Affiliated Members are liable only for the dues and assessments of the
Local Sections.

Contributing Members shall pay no initiation fee. The annual dues shall
be fixed by the Board of Trustees, but shall not be less than twenty-five
dollars, four dollars of which shall be a subscription to the Journal of the
American Ceramic Society. The privileges of membership shall begin upon
payment of the annual dues.

Any Active or Associate Member in arrears for over one year may be sus-
pended from membership by the Board of Trustees until such arrears are

AMERICAN CERAMIC SOCIETY. 857

paid and in event of continued dereliction, may be dropped from the rolls.
Active Members in arrears are not eligible to vote. The annual dues of
Active and Associate Members are payable within three months succeeding
the date of the annual meeting.

IV.
OFFICERS.

The affairs of the Society shall be managed by a Board of Trustees, con-
sisting of a President, Vice-President, Honorary Secretary, Secretary, Trea-
surer, and five Trustees. The officers, other than the Honorary Secretary
and three Trustees, shall be elected from the Active Members at the annual
meeting and shall hold office until their successors are elected and installed.
The retiring President shall serve as Trustee for two years.

The President, Vice-President, Secretary, and Treasurer shall be elected
for one year and the Trustees for three years; and no President, Vice-Presi-
dent, or Trustee shall be eligible for immediate re-election to the same office.

The Board of Trustees may from time to time, within their discretion, elect
any past Secretary of the Society as Honorary Secretary to perform such
duties as they may designate.

The President shall have general supervision of the affairs of the Society
under the direction of the Board of Trustees and shall perform such other
duties as pertain to his office.

The Secretary shall be, under the direction of the President and Board
of Trustees, the executive officer of the Society. He will be expected to at-
tend all meetings of the Society and of the Board of Trustees, prepare the
business therefor and duly record the proceedings thereof.

He shall see that all moneys due the Society are carefully collected and
without loss transferred to the custody of the Treasurer. He shall carefully
scrutinize all expenditures and use his best endeavor to secure economy in
the administration of the Society. He shall personally certify the accuracy
of all bills or vouchers on which money is to be paid and shall countersign
the checks drawn by the Treasurer against the funds of the Society when
such drafts are known by him to be proper and duly authorized by the Board
of Trustees. He shall have charge of the books of account of the Society
and shall furnish monthly to the Board of Trustees a statement of monthly
balances. He shall present annually to the Board of Trustees a balance
sheet of his books as of the 31st of January and shall furnish from time to
time such other statements as may be required of him.

He shall conduct the correspondence of the Society and keep full records
of the same. He shall transmit promptly to the Board of Trustees for their
consideration all communications not of routine nature. He shall report
promptly to all Active Members the results of all balloting on the business
of the Society unless otherwise ordered by the Board of Trustees. He shall
perform all other duties which may from time to time be assigned to him by
the Board of Trustees.

ssx JOURNAL OF THE

The Secretary may be paid a salary to be determined by the Hoard of
Trustees.

The Secretary shall furnish a suitable bond for the satisfactory performam i
of his duties.

The Treasurer shall receive all money and deposit the same- in the name
of the Society He shall invest all funds, not needed for current disburse-
ments, as shall be ordered by the Hoard of Trustees. He shall pay all bills
when eei titicd and audited as provided by these Rules or as ordered by the
Hoard of Trustees. He shall make an annual report and sueh other reports
as may be prescribed by the Board of Trusties He shall furnish a satisfac-
tory bond for the proper performance of his duties.

A vacancy in any office shall be filled by appointment by the Board of
Trustees but the new incumbent shall not thereby be rendered ineligible
for re-election to the same office at the next annual meeting. On the failure
of any officer or any member of a committee to execute his duties within a
reasonable time, the Board of Trustees, after duly warning such person,
may declare the office vacant and appoint a new incumbent.

A majority of the Board of Trustees shall constitute a quorum, but the
Board of Trustees shall be permitted to carry on such business as they may
desire by letter.

ELECTIONS.

At the annual meeting a Nominating Committee of five Active Mem-
bers, not officers of the Society, shall be appointed.

At least sixty days before the annual meeting this committee shall send the
names of the nominees to the Secretary, who will immediately send a copy
of the same to each Active Member. Any five Active Members may act as a
self-constituted Nominating Committee and present the names of any nomi-
nees to the Secretary, provided this is done at least thirty days before the
annual meeting. The names of all nominees, provided their assent has been
obtained, shall be placed on the ballot without distinction as to nomination
by the regular or self-constituted Nominating Committee, and shall be mailed
to each Active Member, not in arrears, at least twenty days before the annual
meeting. The voting shall be confined to the names appearing on this ballot.
The ballot shall be enclosed in an inner blank envelope and the outer envelope
shall be endorsed by the voter and mailed to the Secretary. The blank
envelope shall be opened by three scrutineers appointed by the President, who
will report the result of the election at the last session of the annual meeting.
A plurality of affirmative votes cast shall elect.

VI.
MEETINGS.

The annual meeting shall take place on the first Monday in February, or
as soon thereafter as can be arranged, at such place as the Board of Trustees

AMERICAN CERAMIC SOCIETY.

859

may decide, at which time reports shall be made by the Board of Trustees,
Treasurer, and scrutineers of election, and the accounts of the Treasurer
shall be audited by a committee of three appointed by the President. Ten
active members shall constitute a quorum at any regular meeting and a ma-
jority shall rule unless otherwise specified.

The order of business at the annual meeting shall be:

1 . President's address.

2. Reading of minutes of last meeting.

3. Reports of the Board of Trustees and Treasurer.

4. Old business.

5. New business.

6. Reading of papers.

7. Announcement of election of officers. Honorary and Active

Members.

8. Installation of officers and new members.

9. Appointment of committees.
10. Adjournment.

Other meetings may be held at such times and places during the year
as the Board of Trustees may decide, but at least twenty days' notice shall
be given of any meeting.

The President shall appoint at the annual meeting a committee of five,
to be known as the Summer Meeting Committee, whose duty it shall be to
arrange for a summer excursion meeting at some suitable point. The ex-
penses of the Summer Meeting Committee in arranging the program of
visits and for printing, rooms, and facilities for meetings shall be borne by
the Society.

vn.

STANDING COMMITTEES.

The following Standing Committees shall be appointed annually by the
Board of Trustees:

1. Rules.

2. Publications.

3. Membership.

4. Standards.

5. Local and Student Sections.

6. Papers and Programs.

7. Cooperation.

8. War Service.

The Committee on Rules shall consist of five members. It shall receive
all recommendations relating to changes of Rules and shall report upon the
same to the Secretary for transmission to the Society. It shall have power
to propose changes in the Rules of the Society.

86o JOURNAL OF THE

The Committee on Publications shall consist of the Bditor and foui mem
bers. The duties «>f the Bditoi and ( ommittee on Publications arc defined
uikIii \i tide X I on Publications

The Committee on Membership shall consist of at least five members and
shall have the power to appoint sni> committees. Its function shall be to
undertake systematically the enlargement of the membership of the Society

amongst those interested in the silicate and allied industries

The Committee on Standards shall consist of at least five members and
shall have the power to appoint sub-committees. It shall submit written
reports, resolutions, and recommendations, relating to tests of materials
and products. The Committee may report at any regular meeting of the
Society. For adoption, these reports, resolutions, and recommendations
must be submitted in printed form to the members of the Society at least six
months before a vote may be taken, during which time any amendments,
changes, or corrections suggested by any member may with the approval
of the Committee be incorporated in the report, resolution, or recommenda-
tion. The reports, resolutions, and recommendations as amended shall
then be submitted by letter ballot to the Active Members. A two-thirds
vote shall be required for adoption.

The Committee on Local and Student Sections shall consist of five mem-
bers. Its duties shall be to promote the organization and welfare of Local
and Student Sections.

The Committee on Papers and Programs shall consist of the Chairmen of
the various Divisions, the Secretary of the Society, and such other persons as
the Board of Trustees may deem advisable. Its duties shall be to procure
papers and discussions for the meetings and publications of the Society.
The Committee may require an abstract of any paper submitted before
placing it upon a program.

The Committee on Cooperation shall consist of at least five members.
Its duties shall be to promote the interests of the Society by cooperation
with other societies and organizations and to secure recognition of the So-
ciety by any proper means.

The Committee on War Service shall consist of at least five members.
Its duties shall be to investigate National problems pertaining to the silicate
industries and to stimulate research and production.

VIII.
DIVISIONS.

Professional groups to be known as Divisions of the Society and to be or-
ganized from members of the Society may be authorized by the Board of
Trustees for stimulating the growth and development of the Society, when
such action shall seem wise and expedient. Any member of the Society
may register in any Divisions of the Society in which he is interested.

The officers of a Division shall be a Chairman and a Recording Secretary.
Officers of Divisions shall be elected annually by ballot at the last session

AMERICAN CERAMIC SOCIETY. 86 1

of the Division held during the annual meeting of the Society and shall take
office at the close of the meeting at which they were elected. They shall
hold office for one year or until their successors are elected. The Board of
Trustees shall fill any vacancies occurring through death or resignation
among the officers of a Division.

A Division shall have the right to make by-laws for its own government,
which shall be subject to the approval of the Board of Trustees and must
not be inconsistent with the Rules of the Society, and a committee to draw
up such by-laws shall be appointed at the first meeting held by the Division.

Any Division may raise or collect funds to be expended for its own pur-
poses and may have the entire management and control of said funds, in so
far as said management and control does not conflict with any provisions
of the Rules or with the Charter of the Society.

Any Division may be dissolved by the Board of Trustees for good and
sufficient reasons.

IX.
LOCAL SECTIONS.

Local Sections, each carrying some distinguishing title prefixed to the
words "Section of the American Ceramic Society," may be authorized by
the Society.

The purposes of such sections shall be to strengthen and extend the work
of the Society as defined in Section I of its Rules by more frequent meetings
in local centers than are possible to the Society as a whole and by bringing
the benefits of the work to persons who would not otherwise be reached.

Application for permission to form a Local Section must be in writing
and signed by not less than ten members of the Society in good standing,
residing in the general locality where the Section is to be formed, of whom
one at least shall be an Active Member. To be considered at any given
meeting an application must be filed with the Secretary at least thirty days
prior to the date of the meeting and notice that the application is pending
must appear in the program of the meeting. To be granted the application
must receive the affirmative vote of two-thirds of those present. In event of
affirmative action the Society will issue a charter to the applicants authorizing
them to form a Section under the name proposed. Charters for Local Sec-
tions may be temporarily suspended by the Board of Trustees for cause,
but no charter can be permanently rescinded except by vote of two-thirds of
those present at a regular meeting of the Society, after due publication in
the program of the meeting that the matter is pending.

Local Sections shall have power to make their own rules and by-laws
except that they shall not pass any rule or by-law which is in conflict with
the Rules of the Society.

The officers of Local Sections shall be a Chairman, a Secretary, a Coun-
cilor, and such others as the Section may prescribe. The duties of the Chair-
man and Secretary shall be such as usually pertain to those offices. The

S6a JOURNAL I >l; THE

Councilor shall be an Active Member of the Society. He shall be Hwted bj

the Local Section and it shall he his duty to advise the Section in all matt I
pertaining to its relations with the Society and to make an annual report
to the Society regarding the work and status of the Section. The names of
the Chairman, Secretary, and Councilor of each Section shall appear in the
roster of the Society.

Any person interested in the silicate and allied industries is eligible to
membership in a Local Section.

Local Sections shall have power to fix their own dues or assessments, such
dues or assessments being in addition to and independent of the regular
dues and assessments of the Society upon its members. No Section shall
have authority to incur debt in the name of the Society or for which the .So-
ciety may become liable.

X.

STUDENT BRANCHES.

Student Branches, each carrying some distinguishing title prefixed to the
words "Student Branch of the American Ceramic Society," may be estab-
lished in institutions in which regular courses of instruction in ceramics are
maintained.

The purposes of such Student Branches shall be to strengthen and extend
the work of the Society, as defined in Section I of the Rules, by enlisting the
interest and support of students in ceramics while still in school and by stim-
ulating the spirit of ceramic research among them.

Application to form a Student Branch in any institution must be in writing,
signed by not less than five regularly enrolled students in good standing
and endorsed by two or more members of the Society. The application must
be filed and acted upon as provided for Local Sections in Section IX, and
may be suspended or revoked for cause in the same manner.

Student Branches shall have power to make their own rules and by-laws,
except that they shall not pass any rule or by-law in conflict with the Rules
of the Society.

The officers of Student Branches shall be a Chairman, a Secretary, a Coun-
cilor, and such others as the Student Branch may prescribe. The Chairman
and Secretary shall be elected by the Student Branch and their duties shall
be such as usually pertain to those offices. The Councilor shall be an Active
Member of the Society, appointed by the Board of Trustees to act in this
capacity to the Student Branch. The duties of the Councilor shall be to
advise the Student Branch in all matters pertaining to its relations to the
Society and to make an annual report to the Society regarding the work
and status of the Student Branch. The names of the Chairman, Secretary,
and Councilor for each Student Branch shall appear in the roster of the So-
ciety.

Membership in a Student Branch shall be limited to the instructional
force and regularly enrolled students of schools in which branches are located,

AMERICAN CERAMIC SOCIETY. 863

subject to the rules of the school concerned governing outside activities
of students.

Student Branches shall have power to fix their own dues and assessments
and, as such, shall pay no dues or initiation fees to the Society. No Student
Branch shall have authority to incur debt in the name of the Society or for
which the Society may become liable.

Members of a Student Branch as such shall pay no dues or initiation fees
to the Society but upon depositing a certificate of good standing from their
Secretary, may purchase from the Society its publications at the same rate
as Associate Members of the Society. This privilege shall cease when the
student's connection with the school ceases, but members of a Student Branch,
upon leaving school, may at once become Associate Members of the Society
by depositing their certificates, making proper application, and paying the
regular initiation fee.

XI.
PUBLICATIONS.

The Board of Trustees may employ, at a suitable compensation, an editor
of the publications of the Society.

The Committee on Publications shall have general supervision of the pub-
lications of the Society and of contracts and expenditures connected there-
with, subject to the approval of the Board of Trustees.

In the consideration of papers offered for presentation or publication those
papers containing matter readily found elsewhere, those specially advocating
personal interests, those carelessly prepared or controverting established
facts, and those purely speculative or foreign to the purposes of the Society,
shall be rejected. The Committee on Publications shall determine which
papers shall be printed. The Committee may return a paper to the writer
for correction and emendation and may call to its aid one or more members
of special experience relating to the subject treated, either to advise on the
paper or to discuss it.

The Committee on Publications shall provide for the publication of a
monthly periodical entitled the Journal of the American Ceramic Society, the
subscription price of which shall be four dollars to members and six dollars
to non-members.

One copy of the paper-bound edition of the Journal shall be sent prepaid
to each member of the Society not in arrears. No member shall be furnished
with more than one copy for any single year in return for his regular subscrip-
tion. A member shall be permitted to complete by purchase one file of the
publications of the Society at less than the current commercial rate, the
amount to be fixed by the Board of Trustees and to be called the member's
rate.

The Secretary shall have the custody of all publications of the Society,
shall keep them safely stored and insured, and shall sell them to the public
at prices which shall be fixed by the Board of Trustees. The Board of Trus-

so, JOURNAL OF TIIIv

tees shall also, from time to time, fix the price of the volumes remaining unsold

and shall have authority to refuse to sell the hark volumes of the Trail KU
tions and Journals excepl in complete sets, at such time as the quantity re-
maining of any number becomes so small as iii their judgment to warrant such
action.

The Kditor shall request the author of each article appearing in the Journal
of thi' Society to fill out and sign, within a definite time limit, a form, speci-
fying the number of reprints of said article, if any, which he desires. This
form shall contain a table from which can be computed the approximate
cost at which any reprints will be furnished. In event that the expense of
furnishing the desired number of reprints is large, the Board of Trustees may
require the author to pay part or all of the cost involved before the publica-
tion of the reprints is begun. On receipt of such signed order within the time
limit set the Kditor shall cause to have printed the desired number of copies.
If the author makes no reply or replies after the time limit has expired then
the Society will not be responsible for the publication of any reprints of the
article in question except at the usual market price for the printing of new
matter.

No one shall have the right to demand the publication of an article inde-
pendent of the discussion which accompanied it and no one having taken
part in a discussion upon an article shall be entitled to order reprints of the
discussion separately and apart from the article itself.

The Society is not, as a body, responsible for the statement of facts or
opinions expressed by individuals in its publications.

xn.

PARLIAMENTARY STANDARD.

Roberts' "Rules of Order" shall be the parliamentary standard on all points
not covered by these rules.

xin.

AMENDMENTS.

To amend these Rules the amendment must be presented in writing at a
regular meeting of the Society and, if approved by the Committee on Rules
or by any ten Active Members, must be printed on a ballot and sent out not
earlier than thirty nor later than sixty days after the adjournment of the
meeting at which the amendment was presented. If the said letter ballot
shows an affirmative vote of not less than two-thirds of the total vote cast
within fourteen days of the date of mailing, then the same shall be declared
carried and shall at once become effective.

AMERICAN CERAMIC SOCIETY. 865

PUBLICATIONS.

TRANSACTIONS, AMERICAN CERAMIC SOCIETY.

Bound in
cloth.

Vol. I. 1899, 1 10 pages $4 75

Vol. II. 1900, 278 pages 4 75

Vol. III. 1 901, 238 pages 4 75

Vol. IV 1902, 300 pages 4 75

Vol. V. 1903, 420 pages 5 75

Vol. VI. 1904, 278 pages 4 75

Vol. VII. 1905, 454 pages 4 75

Vol. VIII. 1906, 41 1 pages 4 75

Vol. IX. 1907, 808 pages 5 75

Vol. X. 1908, 582 pages 5 75

Vol. XI. 1909, 632 pages 6 25

Vol. XII. 1910, 882 pages 6 25

Vol. XIII. 1911, 837 pages 6 25

Vol. XIV. 191 2, 888 pages 6 25

Vol. XV. 1913, 747 pages 6 25

Vol. XVI. 1914, 611 pages 6 25

Vol. XVII. 1915, 815 pages 6 25

Vol. XVIII. 1916, 947 pages 6 25

Vol. XIX. 1917, 707 pages 6 25

Complete set $106 75

To members of the Society a reduction of 40 per cent will be made from

the above prices. Members cannot purchase more than one copy of each
volume at members' rate.

NOTE — Vols. II, V, IX, X, XII and XIX will not be sold except in complete sets.

The Society has also published the following books, which will be sold

net, at the prices listed, to the public and members alike:

The Collected Writings of Dr. Hermann August Seger, Volume I,
Contains (a) Treatises of General Scientific Nature, (b) Essays
relating to Brick and Terra Cotta, Earthenware and Stoneware,
and Refractory Wares. 552 pages. Bound in cloth S7 50

The Collected Writings of Dr. Hermann August Seger, Volume II.
(a) Essays on White Ware and Porcelain, (b) Travels, Letters
and Polemics, (c) Uncompleted works and extracts from the

archives of the Royal Factory. 605 pages. Bound in cloth .... 7 50

A Bibliography of Clays and the Ceramic Arts, by Dr. John C. Bran-
ner, 1906. 451 pages. Bound in cloth. Contains 6027 titles

' of works on Ceramic subjects 2 00

A Directory of Dealers in Raw Ceramic Materials, 60 pages 50

866 JOURNAL OF THE

The above publications can be obtained of the Secretary and will be shipped
it the consignee's expense by express or parcels post to any address on re-
ceipl of the price To all who purchase a complete set of the Transactions,
a copy of the Branner Bibliography is given free. All checks or money orders
are payable to The American Ceramic Society.

CHAS. F. BINNS, Secretary.

Alfred, N. Y.

ANNUAL REPORT OF THE BOARD OF TRUSTEES.

in thi: Members ov the American Ceramic Society:

The Board of Trustees has prepared the following statement of the busi-
ness and condition of the Society for the period February i, 1917 to February

1 , 1 9 1 8 :

Receipts.

Brought forward February 1, 1917 $1 ,489.12

Dues $3,642.88

Fees 861 .85

Contributing Members 500 .00

Transactions 2 , 179 .00

Branner 's Bibliography 13 .20

Seger's Collected Writings 75 00

Reprints ' 115 .63

Convention Visitors 19 .00

Error in Deposit of Oct. 26, 1917 2 .00

Total Receipts for year ending January 31, 19 18 $7,408.56 7,408.56

Total $8,897-68

Disbursements.
Publication of Vol. XIX:

Stenographic report $ 154 .00

Telegrams 4 .02

Drawings, Illustrations and Engravings 487 . 74

Ptg., Binding, Casing and Typing 1 ,464.93

Editor's salary 600 .00

2 ,710.69

Reprints of Vol. XVII and XVIII 246 .75

Transportation of volumes 186 . 73

Balance due on Volume XVIII 105 .59

Convention expenses 239 .80

Seger's Collected Writings and Binding 75 . 15

Binding early volumes no .65

AMERICAN CERAMIC SOCIETY. 867

Association Dues (Nat'l Fire Prot. Ass'n) 30.00

Membership Committee • 227 .46

Traveling expenses (arranging Summer meeting) 35.00

Nat'l Exposition Chemical Industry 44 50

Refund 3 .75

Insurance and storage on volumes 88 .75

Salary Ass't Secretary 300 .00

Extra clerk hire 577 .25

Postage, Stationery, Supplies, etc 762 .99

Telegrams 7.82

Typewriter Rental 40 . 00

Exchange charges 1 .00

Total Disbursements for year $5 , 793 . 88

Balance Jan. 31, 1918 $3,013 .80

REPORTS OF COMMITTEES AND STUDENT BRANCHES,
1918 MEETING.

REPORT OF THE COMMITTEE ON MILITARY AND ECONOMIC
PREPAREDNESS.

This committee was created at the New York meeting and Prof.
Edward Orton, Jr., was appointed chairman. Fortunately, he
was able to remain with the work long enough to organize it
before we lost him to the Army. The scope of the work comprised
activities dealing, first, with direct war requirements, and second,
those referring to economic matters affected by the war. The
field was divided into sections according to the products most
concerned with war needs. At the initial meeting the following
sub-committees were organized:

Abrasives, R. C. Purdy, Chairman.
Chemical Porcelain, Charles F. Binns, Chairman.
Chemical Stoneware, R. H. Minton, Chairman.
Enameled Iron, R. D. Landrum, Chairman.
Electrical Porcelain, L. E. Barringer, Chairman.
Optical Glass, C. H. Kerr, Chairman.
Raw Materials, A. S. Watts, Chairman.
Refractories, A. V. Bleininger, Chairman.

The work of these committees for the year 191 7-1 8 may be
briefly summarized as follows:

Abrasives. — This committee, principally through its chairman,
began its activity without delay by securing concerted action on
the part of the manufacturers of grinding wheels. From among
them a traffic committee was at once organized whose duty it
was to move goods through embargoes so that each company
could stock its yards with raw material. This was accomplished.
Another committee was appointed to simplify the specifications
for wheels which brought them to the question of machines and
especially the size of arbors. It was necessary to specify the

AMERICAN CERAMIC SOCIETY. 869

arbor size in order that the Government could secure as many
machines of one type as possible. There are also several other
parts of machines that control the specifications of wheels and
these have been considered by several committees. The Grind-
ing Wheel Manufacturers' Association also has a War Service
Committee, one member of which is in Washington all of the time.
All of the work has been done by this organization but by men
who are also members of the American Ceramic Society. The
work has been completed and the final reports submitted.

Chemical Porcelain. — The principal needs of this work con-
sisted in securing adequate specifications which would insure
the production of good porcelain which would meet the require-
ments of the users. This has been done by the chairman who
has devoted much time and attention to this subject. In a num-
ber of instances the specifications have been used to select ware
for Government use.

Chemical Stoneware. — The work of this committee consisted to
a large part in bringing about an organization of manufacturers
for the purpose of securing a sufficient "supply of clay, fuel and
wood for forms, through the rigid embargoes. This involved a
great deal of work on the part of Mr. Minton and his associates.
They also looked after the prompt shipment of goods to the chemi-
cal plants which, in spite of the urgent need, was sometimes very
difficult to accomplish, especially in the congested districts of
the East. It was necessary to keep in touch with the War In-
dustries Board, the various traffic committees, as well as the mili-
tary authorities concerned. Some work has also been done
towards the standardization of chemical stoneware.

Enameled Iron. — The work of this committee was of a some-
what diverse character. Assistance was rendered the Signal
Corps to secure sheet metal parts through the use of the punch
presses operated by manufacturers of enameled iron. The com-
mittee was able also to help the Government in securing the desired
quantity and quality of enameled steel cooking utensils. Just
now the Committee is working upon a standard specification for
cooking utensils and hospital supplies for the Army and Navy.

n7<> JOURNAL OF THE

This committee was instrumental also in having the- manufac-
turers of sheet steel ware form a War Service Committee.

Electrical Porcelain. This committee compiled a list of the
porcelain articles needed by the Government but its attention
was centered upon the spark plug problem, especially with refer-
ence to aeroplane engines. The several members of this com-
mittee, the laboratory of the General Electric Company and
the Bureau of Standards worked along these lines, with reference
to the spark plug itself, the cementing of the electrode and the
assembling of the parts. Considerable work has been done
along these lines and a general improvement has resulted from
these efforts. The use of porcelains of a much higher maturing
temperature, lower feldspar content and more constant thermal
expansion has been brought about. Such problems are not re-
stricted to the spark plug but apply to magnetos as well. In
cooperation with the English a satisfactory insulation for mag-
netos has been worked out which is being manufactured in England
today.

Optical Glass. — This committee was in touch with the develop-
ments in the production of this important material as well as
with its most efficient use in optical instruments due to the con-
nections of the chairman. The committee supplied technical
information, abstracts, etc., to those interested and rendered
such service as was within its province.

Raw Materials.— There has been a distinct lack of information
concerning the sources of American raw materials used in the
ceramic industries, a condition particularly trying during the
stress of war time. Frequently, owing to embargoes or freight
tie-ups, it has been found necessary to make changes in materials
and for this purpose complete lists of miners and dealers were
urgently needed. This need has been supplied by the committee,
principally through the hard work of its chairman, by the publica-
tion of a Directory giving the names and addresses of 545 dealers
in ceramic materials. This publication is still undergoing re-
vision and will be reprinted from time to time.

AMERICAN CERAMIC SOCIETY. 871

Refractories. — This committee associated itself with the Re-
fractories Manufacturers' Association and it was found desirable
to organize a joint committee consisting of representatives of the
principal government departments and manufacturers. This
organization succeeded in interesting such men as Admiral Dyson,
Commander Penn and other government officials in its work.
Several meetings were held in Washington and the standard
shapes adopted by the Refractories Manufacturers' Association
were submitted for general acceptance by all the government
departments. Work was begun on an extensive series of tests
for the purpose of arriving at specifications covering the quality
of refractories. This study was completed and the results issued
by the Refractories Manufacturers' Association for distribution
among its members. A great deal of work also had to be done
with reference to clays for glass pots, graphite crucibles, special
refractories such as magnesia spinel, and questions constantly
brought up by different government departments. In cooper-
ation with the Navy a study was made of light weight refrac-
tories for torpedo boats, destroyers and cruisers.

A. V. Bleininger, Chairman.
February 11, 1918.

REPORT OF COMMITTEE ON SUMMER MEETING.

The members of this committee began early to canvass for
the most favorable itinerary for a summer meeting. We took
into consideration the fact that our people were interested most
largely in the war and its issues as well as the character of the
itinerary.

We met in Hornell, N. Y., Sunday afternoon and Monday morn-
ing of July 8 and 9, 19 17. Those who met with Prof. Binns
and Prof. Shaw on Sunday afternoon had a very delightful social
time. On Monday, July 9th, 32 members and delegates went
by motor bus from Hornell to Alfred. Here the day was spent
most pleasantly with the faculty of the New York State School
of Ceramics at Alfred University. The hearty welcome and
cordial entertainment given by the Alfred faculty was most en-
joyable. Much was learned by the visitors in the opportunity

872 JOURNAL OF THE

to see in detail the most complete equipment and facilities for
the teaching of Ceramics and for the making of ceramic researches.
There probably is no laboratory more thoroughly equipped and
more thoroughly organized than the laboratories of this school.

bate Monday afternoon the party motored back to Ilornell
and by train went to Corning where we spent a most delightful
evening with Messrs. M. E. Gregory, F. R. Carder, Alfred
Maltby and Dr. E. C. Sullivan.

Tuesday was devoted to visits to the plants of the Corning Brick,
Terra Cotta & Tile Co., Corning Glass Works and the Stuben Glass
Co. The members had the rare privilege of seeing in detail the
manufacture of the now famous Pyrex Glass and had described
and illustrated to them the manufacture of art glassware. The
members had a rare opportunity in these visits to the Corning
and Stuben Glass works and were very enthusiastic in their ex-
pression of appreciation.

Tuesday afternoon the Corning members entertained the visi-
tors at a delightful dinner at the Corning City Club.

Wednesday morning the delegation which had now been in-
creased in numbers took an early train from Corning to Watkins
Glen, where they viewed one of the most picturesque wonders
of Nature in America. Pictures and words would fail to convey
an adequate description of the Glen. All who saw it made pledges
to themselves for an early return visit.

After a delightful stay in and about the Glen and the Park the
members took a noon train to Canandaigua where Clarence C.
Keehan, Secretary and Assistant Manager of the Liske Mfg.
Co., met the party, guided them to a restaurant where lunch
had been arranged, and then through the large manufacturing
plants.

The party arrived in Rochester, N. Y., late Wednesday after-
noon. In the evening the party sought its own entertainment
but it was mostly spent in exchange of greetings with the Rochester
hosts and those who joined the party at this time. Thursday
was spent in visiting the enameling plant of the Pfaudler Co.,
Taylor Instrument Co., and the Pennsylvania Feldspar Plant.
The delegates were delightfully entertained at luncheon as guests

AMERICAN CERAMIC SOCIETY. 873

of these companies, and later at the beautiful park on Lake On-
tario.

The party disbanded Friday evening after having enjoyed a
week that was replete in things of Ceramic interest and oppor-
tunities, of seeing ceramic work which to most was not only novel
but unknown, and seeing some of the most wonderful things that
Nature has builded, with companionship that has already blos-
somed into lasting friendship on the part of many and, most, im-
portant of all, with a better understanding of the accomplishments
and purposes of this Society, not only in general, but also as it
relates to the individual members.

The official roster shows that nearly 60 members of the Society
availed themselves of the opportunity of this meeting at one
time or another and that there was at any time a minimum at-
tendance of 32.

R. C. Purdy, Chairman.
February 12, 19 18.

REPORT OF UNIVERSITY OF ILLINOIS STUDENT BRANCH.

The activities of the Student Branch have been somewhat
curtailed during the past year owing to the generally prevailing
conditions. A number of the members left school to enter mili-
tary service before the end of the year. A further portion entered
the Service during the summer, leaving only a small nucleus of
members who returned to take up the work of the Society this
year. The active membership has thus been reduced to fifteen.

The disorganization general in student bodies and the unsettled
state of mind that most of the boys were in has further contributed
to the /fficulty of carrying out a well-planned program for the
activity of the present year. Under these adverse conditions
the members who have carried on the work this year have done
exceedingly well. The following papers and lectures have been
given at the several meetings:

"Plant Experiences," by J. W. Moncrieff and H. H. Sortwell.
"Firing Kilns with Forced Draft," by B. F. Carter.

874 J( »URNAL OF THE

"Electric Furnaces," by F. F. Footitt.

"Class Notes," by J. E. Seabright.

"The Manufacture' of Architectural Terra Cotta," An illus
trated lecture, by Professor C. W. Parmelee.

"Research Notes," by Professors E. W. Washhurn, C. W. Par-
melee and R. K. Hursh.

"The Manufacture of Abrasives," an illustrated lecture, by
Mr. Hermann A. Plusch, Abrasive Co., Phila., Pa.

One of the meetings during the year was made an "Open House"
evening at the Ceramics building, when friends of the members
and others who were interested were shown the laboratories and
saw something of the work. This was especially planned for the
benefit of the freshmen who had thus far had no opportunity
to learn the character of the work ahead of them.

The Student Branch obtained and presented to the Department
of Ceramic Engineering a large framed portrait of Professor
Charles Wesley Rolfe who was so largely instrumental in ob-
taining the establishment of the Department at the University.

The present officers of the Branch at Illinois are: President;
James W. Moncrieff; Vice-President, Griffith S. Kennelley;
Secretary-Treasurer, Benjamin F. Carter.

R. K. Hursh, Councilor.
February 12, 1918.

REPORT OF THE IOWA STATE COLLEGE STUDENT BRANCH.

The number of students in ceramic engineering at Iowa State
College has decreased very much on account of the War. Of
the small enrollment we had last year, there are now eight men
actually in the Army, three of these being in France. Several
of the men enrolled had been drafted and expected to be called
to the colors at an early date.

Nevertheless, the men remaining have held regular meetings
every two weeks of the school year and have shown considerable
enthusiasm in the work of the Student Branch. The arrange-
ment and presentation of the programs has been entirely in the
hands of the students, the faculty taking no hand in these. The

AMERICAN CERAMIC SOCIETY. 875

young men have shown originality and initiative in this work
and a wide variety of subjects, most of them relating to praet; ;al
questions in ceramic engineering, have been discussed.

Mr. E. Raddewig is Chairman and Mr. F. A. Ohlsan is vSecretary
of the Student Branch.

Homer F. Stalky, Councilor.
February 12, 1918.

REPORT OF THE NEW YORK STATE STUDENT BRANCH.

The Society was greatly interested at the opening of the year
in reports given by some of the members who had spent the Sum-
mer in practical work. Two successive meetings were arranged
at which these reports were presented and discussed. It was
felt to be very desirable that students would find opportunity
during their college course to spend one or more summers in this
way.

At the meeting of November 14, Prof. J. B. Shaw gave a talk
on the production and use of enamels for steel and iron. The
comprehensive and practical view given of this part of the ceramic
field was new to many of the members and a great deal of interest
was aroused.

About this time the Student Branch was somewhat disorganized
and temporarily depressed by the loss of its chairman. This
young man, a senior, had been carrying a heavy burden at home
in the sickness and ultimate death of his father. Being thus com-
pelled to postpone his graduation for a year he left college and
went to work in the industry. After a short period of readjust-
ment a new chairman was elected and the work was resumed.
An interesting meeting resulted from a desire to hear what the
lower classmen could do. A member of the Sophomore class was
placed in charge and a program was successfully carried through.

Following upon this a session was devoted to hearing theses
problems discussed by the senior class and finally an evening was
assigned to a consideration of the preparation and use of the dif-
ferent types of cement.

Reviewing the year as a whole the result of the work has been
satisfactory but not ideal. The dislocation alluded to, though

876 I* 'I KNAI, OF THE

temporary, was serious because the school year is too short to
admit of effective readjustments. Each year, however, pro-
vides a small supply of experience which will gradually accumu-
late into a fund.

Ciias. V. Binns, Councilor.
February 12, 19 18.

REPORT OF OHIO STATE UNIVERSITY STUDENT BRANCH.

The past year has been very satisfactory considering the fact
that about fifteen of our students have enlisted, leaving only
about thirty-five men in the course. We have, however, main-
tained our organization and held the following interesting meet-
ings:

October 23, 19 17. Meeting for organization and election
of the following officers: E. D. Vance, Chairman; C. M. Dodd,
Vice-Chairman; and J. W. Hepplewhite, Secretary-Treasurer.
The object and purposes of the Society were explained by Pro-
fessor Watts.

November 20, 191 7. Address by R. T. Stull, U. S. Bureau
of Mines, on "Paving Brick."

January 15, 19 18. Address by Mr. H. G. Schurecht, U. S.
Bureau of Mines, on "The Manufacture of Glass Pots."

The financial statement of the Branch shows $37.24 on hand,
$48.00 collected, and $84.04 expended by order of the Branch,
leaving a cash balance of $1.20.

Arthur S. Watts, Councilor.
February 12, 1918.

IN RETROSPECT.

The first volume of the Journal of the American Ceramic Society is com-
pleted with this number.

Appointed in February 191 8, the Committee on Publications was
authorized to publish a monthly Journal to take the place of the annual
volume of Transactions. The Committee took up the work promptly but
several months were required to decide upon the many questions involved,
such as the proper title and form of the magazine, actual size, subject matter
to be included, advertising, publishing contracts, copyright, etc., etc. It

AMERICAN CERAMIC SOCIETY. 877

was decided to issue the entire twelve numbers for 191 8 even though several
months of the year had already elapsed and after the necessary preliminary
arrangements the January number or Volume 1, No. 1, appeared in July,
1918.

With but six months in which to issue twelve numbers, considerable exertion
was necessary upon the part of the Editor, the Publishers and all concerned
and many seemingly unavoidable delays were encountered. The volume
has been completed in time, however, to permit of Volume II starting suffi-
ciently early in 1919 to ensure its appearance at approximately equal monthly
intervals and completion by January 1920.

Bearing in mind the broad field of industrial operations covered by the
ceramic industries and the need of the Journal's readers for a well-diversified
list of professional papers, the Committee on Publications has endeavored
to secure meritorious papers in all fields and the following tabulation will
indicate the degree of success attained:

»
Subject. No. of Papers.

Clays — Occurrence and Uses 5

Methods of Testing (raw materials and finished products, physi-
cal and chemical) 10

Drying and Driers 3

Kilns 5

Glass ' 9

Refractories (fire brick, saggers, glass pots, magnesia, etc.) 7

Enamels for cast iron and sheet steel 7

Pottery and porcelain 7

Terra Cotta 4

Cements, lime and plaster 5

Certainly a wide field is covered in this list and there is much of interest
and value to all interested or engaged in ceramic work.

Perhaps some important fields were not sufficiently represented but this is
due entirely to not having contributions from capable men in such branches.
For Volume II the Committee on Publications is in a position to promise even
a greater degree of diversification, while still maintaining and even advancing
the degree of technical excellence.

The Committee on Publications desires to thank all who have contributed
to the success of Volume I, the authors of the various papers published, the
pioneer advertisers who accepted the value of the advertising medium upon
faith, and the members of the American Ceramic Society who have offered
helpful suggestions and cooperation.

We are encouraged by Volume I ; we hope it has proven both pleasing and
profitable to its readers and we will endeavor to serve you as well or better in
Volume II.

Committee on Publications.

TECHNICAL PAPERS VOLUME I.
JOURNAL OF THE AMERICAN CERAMIC SOCIETY.

No. 1.

i . Kaolin in Quebec 8

By J. Keele.

2 . Special Pots for the Melting of Optical Glass 15

By A. V. Bleininger.

#

3. The Effect of Gravitation upon the Drying of Ceramic Ware 25

By E. W. Washburn.

4. Test of a Producer Gas-Fired Periodic Kiln 35

By C. B. Harrop.

5. Notes on the Hydration of Anhydrite and Dead-Burned Gypsum . . 65

By A. C. Gill.

No. 2.

6. On the Relation between the Physical Properties and Chemical

Composition of Glass: VIII. Molecular Compounds 76

By E. W. Tillotson, Jr.

7. Fire Clays in Northern Idaho 94

By E. K. Soper.

8. Ground Coat Enamels for Cast Iron 99

By Homer F. Staley.

9. The Bonding Strengths of a Number of Clays between Normal

Temperature and Red Heat 113

By C. W. Saxe and O. S. Buckner.

10. Observations on the Formation of Seed in Optical Glass Melts. ... 134

By A. E. Williams.

No. 3.

11. Some Phenomena in Glaze Reduction 148

By H. B. Henderson.

12. Natural Hydraulic Cements in New York 160

By Robert W. Jones.

AMERICAN CERAMIC SOCIETY 879

13. Tests of Clays and Limes by the Bureau of Standards Plasticimeter 170

By F. A. Kirkpatrick and W. B. Orange.

14. The Use of Furnace Slag as Grog in Architectural Terra Cotta

Bodies 185

By R. H. Minton.

15. The Effect of Electrolytes on Some Properties of Clays 201

By H. G. Schurecht.

No. 4.

16. The Effect of the Degree of Smelting on the Properties of a Frit. . 221

By E. P. Poste and B. A. Rice.

17. Cobalt-Uranium Green Glaze for Terra Cotta 238

By Hewitt Wilson.

18. Alabaster Glass: History and Composition 247

By Alexander Silverman.

19. The Zwermann Tunnel Kiln and Its Operation 262

By Carl H. Zwermann.

20. The Properties of Some Ohio and Pennsylvania Stoneware Clays. . 267

By H. G. Schurecht.

2 1 . The Use of Dense Solutions in Determining the Structure of Porce-

lain 287

By H. Spurrier.

No. 5.

22. A Study of Draft Movements in Flues 294

By E. H. Fritz.

23. The Clays of Florida 313

By E. H. Sellards.

24. An Attempted Heat Balance on a Continuous Kiln of the Chamber

Type 322

By Chester Treischel and H. S. Robertson.

25. Notes on the Laboratory Testing of Silica Brick 338

By R. J. Montgomery and L. R. Office.

26. Polychrome Decoration of Terra Cotta with Soluble Metallic Salts . 353
By Hewitt Wilson.

No. 6.

27. The Porosity and Volume Changes of Clay Fire Brick at Furnace
Temperatures 384

By G. A. Loomis.

880 l« TKXAI. OF THE

28 The Calculation of the "Rational Analysis" of Clays 404

By Henry S. Washington.

29 The Action o!' Acetic Acid Solutions of Different Strengths on a

Sheet Steel Enamel 422

By I. J Frost.

30. Note on Colors Produced by the Use of Soluble Metallic- Salts

Bj Paul G. Larkin.

31, Note on the Sintering of Magnesia

By J. B. Ferguson.

32. The Occurrence of High-Grade American Clays and the Possi

bilitics of their Further Development 446

By H. Ries.

33. The Effect of Certain Impurities in Causing Milkiness in Optical

Glass 4"8

By C. N. Fenner and J. B. Ferguson.

34. Silica Refractories 477

By Donald W. Ross.

35. Antimony Oxide as an Opaciher in Cast Iron luiamels 502

By J. B. Shaw.

36. Newcomb Pottery 5 '8

By Mary G. Sheerer and Paul E. Cox.

37. A Conveyor Stove Room 529

By C. E. Jackson.

38. Preparation and Application of Enamels for Cast Iron 534

By Homer F. Staley.

39. A Pycnometer Operated as a Volumeter 55^

By H. G. Schurecht.

40. Communicated Discussion by F. Gelstharp on the paper "The

Effect of Certain Impurities in Causing Milkiness in Optical
Glass," by Fenner and Ferguson 55S

41. Heat Balance on a Producer-Gas Fired Chamber Kiln 567

By R. K. Hursh.

42. Some Factors Influencing the Time of Set of Calcined Gypsum 57*

By F. F. Householder

AMERICAN CERAMIC SOCIETY. 88 1

43. The Nature of the Air Content of Pugged Clays 584

By H. Spurrier.

No. 9.

44. The Identification of "Stones" in Glass .594

By N. L. Bowen.

45. Some Types of Porcelain 606

By F. H. Riddle and W. W. McDanel.

46. The Use of Car Tunnel Kilns for Brick and Other Products of Crude

Clays 628

By Ellis Lovejoy.

47. The Presence of Iron in the Furnace Atmosphere as a Source of

Color in the Manufacture of Optical Glass 637

By E. W. Washburn.

48. The Control of the Luster of Enamels 640

By Homer F. Staley.

49. observations on Apparent Causes of Failure of Lead Glass Pots. . 648

By A. F. Gorton.

50. An Unusual Cause of Spalling of Sewer Pipe 660

By C. W. Parmelee.

No. 10.

5 1 . A I Hscussion of Terra-Cotta Dryers 667

By Ellis Lovejoy.

52 The Composition of Chinese Celadon Pottery 675

By J. S. Laird.

53. The Hydraulic Properties of the Calcium Aluminates 679

By P. H. Bates.

54. Note on Certain Characteristics of Porcelain 697

By A. V. Bleininger.

.55. Enamels for Cast Iron 703

By Homer F. Staley.

56. A Method for the Determination of Air in Plastic Clay 710

By H. Spurrier.

57. Notes on Sagger Clays and Mixtures 716

By G. H. Brown.

58. Magnesia Ware 730

By Trygve D. Yensen.

882 AMERICAN CERAMIC SOCIETY.

No. 11.

,sg. A Contribution tO the Methods of Class Analysis, with Special

Reference tO Boric Add and the Two Oxides of Arsenic 739.

By E T Allen and ]•;. G. Zies.

oo The Condition of Arsenic in Glass and Its Role in Glass-Making. . 787
By Iv. T. Allen and B. G. Zies.

61. The Partial Purification of Zirconium Oxide 791

By A. J. Phillips.

62. Strength Tests of Plain and Protective Sheet Glass 801

By T. L. Sorey.

INDEX TO VOLUME I.

JOURNAL OF THE AMERICAN CERAMIC SOCIETY.

A

Absorption of terra cotta bodies 187

Acetic acid, Effect of, on sheet steel enamels.

Frost 422

Acid test on enamel frits 223

Ageing, Effect of electrolytes in 214

Effect of, on viscosity of porcelain

bodies 179

Effect of, on porcelain bodies 616

Agricultural products, statistics, compara-
tive 376

Air, in pugged clays. Spurrier 584

I in plastic clay, Determination of.

Spurrier 710

Alabaster glass, History and composition of.

Silverman 247

from Hillsboro, N. B 65

Alabama kaolins, Occurrence of 452

.Albery, D. F., Discussion of green terra

cotta glazes 243

Albite-anorthite glasses, Refractive indices

and densities of 79

Alcohol, used with underglaze colors 354

Wood, Effect of impurities in, on B2O3

determination 766

Alkalies, Effect of, in producing milkiness

in glass 560

Allen, E. T. and Zies, E. G. "Methods of
Glass Analysis with Special Refer-
ence to Boric Acid and the Two

Oxides of Arsenic" 739

-"Condition of Arsenic in Glass and Its

R61e in Glass Making" 787

Alum, Use of, as an accelerator in gypsum 71

Alumina, Effect of, in alabaster glass 252

-Effect of, in silica brick 342

-Effect of, in optical glass 137

-Separation of, from barium and calcium

in glass 779

-Function of, in portland cement 681

Aluminates, calcium, Hydraulic properties

of. Bates 679

AMERICAN CERAMIC SOCIETY:

Annual meeting, 1918 662

Honorary Secretaryship 664

New Members during Sept., 1918 441

New Members during Oct., 1918 514

New Members during Nov., 1918 588

New Members during Dec, 1918 734

Professional Divisions 586

Professional Divisions, Organization

leaders of 663

Reprints from Journal 586

American clays, High-grade. Ries 446

-optical glasses, Analyses of 784

ANALYSES:

-Ashes, from gas producer 330

-Bonding clays used by Saxe and

Buckner 115

ANALYSES (continued):

CO2 in optical glass bubbles 135

Coal charged to gas-producer 330

Chinese Celadon glazes (}7h

Chinese Celadon bodies 677

Connecticut feldspar 504

Discussion of value for fire brick 386

Enamel frits. Poste and Rice 225

Experimental batches for cement!

Bates 6.84

Fused magnesia 732

Flue-gases of producer-gas fired kiln 54

Gases from pugged clay. Spurrier. . . . 585

Limits for silica brick 497

Methods of analyses of boric acid and

the two oxides of arsenic. Allen

and Zies 739

Optical glasses, American 784

Phonalite 368 —

■ —Potashes for glass batches 249

Producer-gas samples 47

Quebec kaolin 10

Rational analysis of clays, Washington 405

Russian potash 469

Searchlight glass 475

Slag from Coatesville, Pa 186

Slag from Illinois Steel Co 189

SO3 and CI in optical glass 473

Typical, of silica brick 347

Uranium oxide 239

Ultimate, value of, for clays. Wash-
ington 409

Anhydrite, and dead-burned gypsum. Gill 65
Anorthite-albite glasses, refractive indices

and densities of 79

Antimony oxide in cast iron enamels.

Shaw . 502

Separation of, from arsenic in analysis

of glass 743

enamels for cast iron 706

APPARATUS:

Conveyor stove room for potteries.

Jackson 529

Expansion of silica brick. Ross 495

Heat balance on continuous kiln 323

Heat balance, description of 42

Measuring draft movements in flues. . . 296

Measuring air content of clays 584

Quantitative determination of air in

clay. Spurrier 712

Tensile strength machine for hot bri-
quettes 117

Application of ground coats to cast iron .... 543

of cover coats to cast iron 552

Arsenic, Effect of, in causing milkiness. . . . 560

in glass, Methods of analysis of. Allen

and Zies 739

Arkansas bond clays, Occurrence of 459

Art pottery, History of Newcomb 518

Atlantic Coastal Plain Belt, Clays in 44g

Aventurine glazes, discussion of I57

883

NNj

T( IURNAL OF THE

B

Ball clays, Tenn and Kv . Viscosity of slips 178

i >, . in renct ■•! in \ mei ica Ries 1 w>

Effect "I ni poi 1 1 lain bodic 6 1 9

H. ill test foi testing fire brick 390

Barium, and lead, Determination <>f. in glass 777

disilicate in optical glass 60
Effect of addition of, to porcelain
bodies

in optical glass 136

Barringer I. E Discussion of effect of

gravity in drying 32

Discussion of -iii in clay 714

Discussion of car tunnel kilns 634

Basalt, Formation of clays from °4

Batch weight, of sillimanite calcine 624

of cone 1 I porcelain body 625

enamels for cast iron .505

Hatch stones in glass 597

Hath tul>s. Enameling of 543

Batting out machine used with jigger. . 532
Hates. l\ II "Hydraulic Properties of the

Calcium Aluminatcs" 679

Bauscfa X: I.omh Optical Co., Manufacture

of optical glass at 468

Beecher, M. P. Discussion of bonding

strength of clays at red heat. ...... 132

Bending tests on sheet and plate glass. . . . 807

Bertrand and Agulhon. Method of testing

for B>< >.. 764

Bibliography on alabaster glass, books 256

on alabaster glass, journals 258

on Fla. clays 321

Binns.C.F. Discussion of glaze reduction 156

Discussion of air in clay 714

Biotite in Idaho granite 96

Bisque, Whiteware. as grog in glass pots. . . 17
Hleininger, A. V. "Certain Characteristics

of Porcelain" 697

"Special Pots for the Melting of

Optical Glass" 15

Blending, of ground coats for enamels 104

Blower used with gas-producer 37

Blue colors, soluble, for terra cotta. Larkin 434

Body mixtures, Newcomb Pottery 522

Porcelain, some types of. Riddle and

McDanel 606

Porcelain, Formulas of 607

of Chinese pottery 678

-terra cotta, Use of slag as grog in.

Mint on 185

Boiling during fritting of enamels 509

Bonding, clays, used by Saxe and Buckner. 115

clays for saggers 719

clays in glass pot mixtures 16

strength of clays dried at different tem-
peratures 119

strengths of clays between normal

temperature and red heat 113

Boric acid, in glass, Methods of analysis

of. Allen and Zies 761

Addition of, to sillimanite calcine 623

oxide, Effect of, on viscosity of enamels 641

Boro-silicate crown optical glass, Milkiness

in 559

Bowen, N. L. "Identification of Stones in

Glass" 594

Bricks, Burning of, in car tunnel kilns 628

Silica. Ross 477

building. Burning, in continuous kiln. . 323

building. Production of. in Fla 313

Brown colors, soluble, blends for terra cotta.

Larkin 435

Brown, <', u •■•., . , , . in,| NIiv

tuns" |,

Bubbll . Ml ;■] , S(;,j

in optical glass in. Il | ^

Buclcnei , O. S and Saxe, ( W Bondini

Strengths of Claj |l]
Bureau of Standard T< i <■,, pia

Clays"

n port on terra cotta t. I I9J

Burning art pottery at Newcomb pottery 527

heavy clay products in car tunnel

kilns ,,.>*
flashed brick 152

natural hydraulic cements .... 161

of silica brick . \i,

— properties of clays, Effect of electro-
lytes on 211

preliminary burning of glass pots. .23. 654

porcelain bodies for glass pots . 610

silica brick, effect of reheating IKfl

strength of clays Saxe and Buckner 113

temperatures of cements 084

Burners, cast iron, used with producer-ga

clay, for glass melting furnaces 639

Oxidation of. in glass melting furnaces M8

Hurt. S G, Discussion of terra cotta glazi

By-product coke oven. Silica brick from }8<>

c

Calcium aluminatcs 079
carbonate as flux in chemical stone-
ware 267

magnesium-silicate glasses. Refractive

indices and densities of 79

Separation of, from alumina and barium

in glass analysis 779

Calcined alumina-clay, Addition of, to

porcelain bodies 623

gypsum, Time of set of 578

kaolin in glass pot mixtures 16

CALCULATIONS

Heat balance on producer-gas fired

periodic kiln 45

Heat balance of continuous kiln 330

Heat balance on chamber kiln 569

Rational analysis of clays. Washing-
ton 405

used with pyenometer- volumeter 558

Canadian China Clay Co. Clav Properties

of ." 8

Carborundum, as coating for chemical

stoneware 268

Carbon in salt glazes 149

Car tunnel kilns. Use of, in burning heavy

clay products 628

Carbon-dioxide in glass, Determination of. . 781

Casting glass pots 19

Cast iron enamels. Preparation and appli-
cation of. Staley 534

enamels, Use of antimony oxide in.

Shaw 502

Ground coat enamels for. Staley. ... 99

enamels. Staley 703

Celluloid coatings on plate and sheet glass 801

Celadon pottery. Laird 675

Cement, portland, Compounds in 679

Natural hydraulic in N. Y. Jones. . 160

Centrifugal force, Effect of, upon drying 33

Ceramic products, statistics, comparative. . 372

industries, classified list 380

Chamber kiln, Heat balance on 567

Chapin's method for determination of boric

acid in glass 762

i

AMERICAN CERAMIC SOCIETY.

885

Chemical compounds in glass 76

glassware, Effect of boric acid in 763

stoneware. Properties of 267

analyses (See Analyses).

Chilling of optical glass melts 135

China clay washing plant, St. Remi,

Quebec 10

bodies, commercial, Properties of 617

Chinese pottery. Laird 675

Chlorides in optical glass 473

zirconium, Formation of 794

Chlorine, Effect of presence of, in glass

analysis 755

Use of, in purification of zirconia . . . . 794

Clays, Air content in pugged 584

American, in porcelain bodies 608

■ Biloxi, Miss., used for saggers 527

content, Effect of, in porcelain bodies. . 611

Effect of electrolytes on properties of. . 201

Function of, in cast iron enamels 101

for saggers. Brown 716

Gcodes in 660

in sheet steel enamels 422

in Northern Idaho 94

of Florida 313

Occurrence of high-grade, in America.

Ries 446

plastic, Determination of, air in 710

Sandy, in terra cotta bodies 197

Strengths of, at red heat. Saxe and

Buckner 113

stoneware, Properties of Ohio and

Penna. Schurecht 267

• Tests of, with plasticimeter 170

Clare, R. L. Discussion of green terra

cotta glazes 243

Theory of, on failure of terra cotta. ... 185

Clinker, cement, Crystallization of 682

Coal Measures, Clays in 464

Coatings for sheet and plate glass. Sorey . . 801

Cobalt stain, Use of, in porcelain bodies. . . . 627

uranium green glaze for terra cotta.

Wilson 238

oxide in cast iron enamels 101

Coke, petroleum, Use of, in chlorination . . . 784

Combustion in car tunnel kilns 629

Columbia basalt in Idaho 94

Color, of enamels, Effect of antimony

oxide on 503

of salt glazes 148

of optical glass from iron in gases. . . . 637

of porcelain, Effect of firing on 626

soluble, for terra cotta. I.arkin 429

underglaze, Formulas of 554

OMMITTEES:

Chemistry of cements and related
building materials National Re-
search Council 5

Chemistry of glass. National Re-
search Council 5

Chemistry of ceramics. National Re-
search Council 5

Standards, A. C. S. Tests of clays. ... 132

oncretc, Compressive strengths of special. 692

ontinuous kiln, Heat balance of 322

fired with producer-gas 35

onveyors, Stove room. Jackson 529

onnecticut kaolins 461

onstants, Table of, for pressure-tempera-
ture conversion of gases 304

onduct.ivity, electrical porcelains 699

of silica brick 343

ooling optical glass 1 34

of terra cotta after firing 185

— cast iron enameled ware 554

Cooling of saggers, Effect of rate of 717

Corundum, used in chemical stoneware. . . . 268

Cornwall, England, china clay 13

Corrosion of glass pots 1 5-596

of glass pots. Discussion of 651

Cost of soluble colors for terra cotta 437

Creighton, E. E. F. Discussion of specific-
gravity test for porcelains 290

Cox, Paul E. "Technical Practice at New-
comb Pottery" 521

Cristobalite crystals in optical glass 475

Properties of 477

Optical properties of 604

Cross-breaking tests on clays 202

Crown drops in glass 598

Crown and boro-silicatc optical glasses. . 134

Crucibles. Clays for graphite 447

-Magnesia 730

Crushing strength of silica brick ^41

test for fire brick 390

Crystallization of gypsum 66

of glasses and enamels 640

of fused magnesia 733

of pot stones 596

Cryolite, Use of, in cast iron enamels 504

Crystalline forms of silica 345

D

Danielson, R. R. Discussion of acetic

acid tests for enamels 427

Discussion of enamel frits 234

Day, Arthur L. Committee on Glass 5

Davenport, R. W. Discussion of draft

movements in flues 308

Dead-burned gypsum, Hydration of. Gill. 65

Deflocculation of clays, Discussion of 214

Decolorizers, Use of, in optical glass 637

Deformation of glass pots, Causes of 656

tests on enamel frits 222

Discussion of, of fire brick 389

Dehydration of gypsum 69

curves, Data for 118

of clay 574

Density of clays, Effect of electrolytes on . 208

of lime pastes, Effect of water on 117

Design of car tunnel kilns 629

Devitrification of optical glass 475

stones in glass 599

Diealcium-silicate, Discussion of 679

Di-electiic strength of porcelains. Effect

of heating on 699

Differential gages on kilns 310

Dolomite, Effect of addition of. to porcelain

bodies 623

Draft movements in flues. Fritz 294

Dredges for applying enamels on cast iron . 553

Dryers for terra cotta. Lovejoy 667

for enamel flits 540

Drying, Effect of, on bonding strengths 1 13

■ -glass pots 20

Effect of gravitation upon the drying

of ceramic ware. Washburn 25

of pottery in conveyor stove room 532

Dunting of terra cotta 1 85

Dusting of portland cement clinker 687

EDITORIALS:

Acid Resisting Enamelware 74

American Ceramic Society Exhibit at

the Exposition 219

American Ceramic Society and the

War 292

I

JOURNAL OF THE

EDITOR] \l.s (continued):

Annual Meeting 515

Chemical Stoneware 291
Products it the Chemical Kx-

tion 167

Co Operation 145

I tirectoi j ni I dealers 517

Puel Curtailment < >rders 2

Growth of t!ii' Society 737

Glassware Exhibits it the Exposition. . 169

High temperature Furnaces 146

imported Clays 665

Industrial Furnace Section 146

Kilns 738

I i«)k in,. Foi " .ird 443

Membership Campaign 147

National Research Council 4

Patent Situation 591

Pottery Industry 666

Professional Divisions 444

Technical Ceramic Training 217

To t he Public 1

To Members 516

-Unity in the Ceramic Industry 73

War Curriculum in Ceramic Engi-
neering 293

Bdgai plastic clays, Occurrence of 318

Efficiency of continuous kilns 333

of producer gas-fired chamber kiln 576

Electrolytes, Effect of, on clays 201

Electric furnace, Sintering magnesia in 440

Electrometric method for determination of

boric acid in glass 762

Electrolytic determination of zinc in optical

glass 775

Electrical porcelain. Firing of, in tunnel

kiln 634

Embayment area. Clays in 453

Emley, W. E Discussion of hydration of

anhydrite and gypsum 70

-Discussion of Plasticimeter 183

Enamels, Ground coat, for cast iron 99

cast iron, Preparation and application

of. Staley 534

Testing sheet steel, with acetic acid.

Frost 422

Luster of. Staley 640

for cast iron. Staley 703

Use of antimony oxide in. Shaw 502

English clays, Impoitation of 447

Engobe. Clay content of, for terra cotta. . . 429

Erie Canal, Use of natural cements in 162

Erosion of glass pots 653

European porcelains, Types of 606

porcelains, Firing of 626

Expansion of silica brick 34 1

of silica brick under load 494

F

Failure of glass pots 648

of terra cotta bodies 185

Fenner, C. N. and Ferguson, J. B. "Dis-
cussion of Milkiness in Optical

Glass" 559

and Ferguson, J. B. "Effect of Cer-
tain Impurities in Causing Milki-
ness in Optical Glass" 468

Feldspar, in glass pot mixtures 16

in Idaho granite 96

■ -as flux in chemical stoneware 267

content, Kffect of, on porcelain bodies . . 611

in clays. Rational analysis of 407

porcelain, Electrical conductivity of. . . 700
Use of, in glass pot mixtures 657

DO, I B "Note on the Sintering

of Magnesia"

and C N Femur Milklni

optical glass 16H,

Ferrous iron, Presence of, iu arsenil .

in Chinese glazi

Ferric oxide in glass furnai i it mo plicre. . .
Pining of optical glass

jfl ii .inn III

Fire clays in northern Idaho. Soper

Fire-cracking of terra cotta

Fire brick, manufacture in Idaho

Porosity and volume changes of, at

furnace temperatures

Effect of iron content in, on glass color. .

-for glass melting furnaces

Filter press leaves, Air in.

Piling, Schedule used In testing fire brick. .

tests, with rifle on sheet and plate

glass

Flashed brick, Color of

Flint content. Effect of. on porcelain
bodies

Flint clays of Missouri, Occurrence of

Floating enamels, Use of magnesium sul-
phate for

Floor dryer for terra cotta

Flotation of cast iron enamels

Florida, Clays of

kaolin, Occurrence of

Flues, for enameling furnaces

Fluxing oxides, Effect of, on luster of
enamels

FORMULAS:

Alabaster glasses. Silverman

Antimony oxide enamels. Staley

Blue stain for terra cotta

Bright glaze for terra cotta

Cast iron enamels. Shaw

Enamel frits

for sheet steel. Frost

Ground coats for cast iron enamels. . . .

Matt glaze for terra cotta. Wilson. . .

Matt glaze, Newcomb Pottery

Normal glass

Purdy's porosity

resistance, Electrical, of porcelain

Refractive index of glass

Shrinkage in terms of true clay volume.

Terra cotta bright glaze. Larkin

Terra cotta matt glaze. Larkin

Terra cotta bodies

— — Tin glaze, Newcomb Pottery

-Tin enamels. Staley

Velocity of gases in flues

Vapor pressure in drying

Forsterite, in sintered magnesia

Frink, R. L. Discussion of seeds in optical
glass

Frit, used in ground coat enamels

Smelting of enamel

Enamel, for cast iron

Fritz, E. H. "A Study of Draft Move-
ments in Flues"

Freezing tests on terra cotta

French flint, in porcelain

Frost, Leon J. "Action of Acetic Acid
Solutions of Different Strengths
on a Sheet Steel Enamel". . . . 422,

Fulton, C. E. Discussion of pots for
optical glass

Discussion of bonding strength

Fuller's earth in Florida

Fuel used in enameling furnaces

7S2

67 7
',38
I 44
788

94
185

•,-

384
638
639
713
393

801
152

612

465

99

673
102
313
450
545

643

247
709
2 }9
241
505
J J 4
422
106
239
521

1 18
700

91

2 05
432
432
18(,
521
706

504

2h
439

l.V.

9'.

221

53(

is;

69t

42:

2'

33!
51
14-'

AMERICAN CERAMIC SOCIETY.

887

Furnaces, Frit, for cast iron enamels 536

• for purification of zirconia 796

Melting, for cast iron enamels 545

atmosphere, Iron in glass 637

Fusibility of cast iron enamels 105

Fused magnesia. Preparation of 730

Fusion point (see Softening Point-).

Gages, Draft. Fritz 294

Ganister rock for silica brick manufacture. . 338
Gates, W. D. Discussion of seeds in optical

glass 136

Gas-producer, Description of 35

' used on chambei kilns 561 ',

Gases, in bubbles in glass 135

waste, Heat losses in 332

Nature of, in clays 584

in glass, Determination of 781

Cclntharp, P. Discussion of seeds in optical

glass 138

Discussion of the "Effect of impurities

in causing milkiness in optical

glass" 559

! Arsenic in glass -.- 787

GEOLOGY:

of northern Idaho fire clays 94

Hydraulic limestones in N, Y 165

Occurrence of Penna. and Ohio stone-
ware clays 269

Occurrence of Fla. clays 313

h Occurrence of quartzites 480

I Occurrence of high-grade American

clays 446

Geodes, in sewer pipe clays 660

Georgia kaolin, Occurrence of 450

German potash in optical glass 469

(Gill, A. C. "Hydration of Anhydrite and

Dead-burned Gypsum" 65

Glaciation, in Canada, Discussion of 13

Glass pots, Special, for the melting of

optica! glass. Bleininger IS

I Seeds in optical. Williams 134

I alabaster, History and composition of.

Silverman 247

— — Identification of stones in. Bowen. ... 594

optical, Milkiness in. Gelstharp 559

-Effect of certain impurities in causing
milkiness in optical. Fenner and

Ferguson 468

pots, Clays for 447

-Methods of analyses of boric acid and

the two oxides of aisenic 739

-optical, R61e of arsenic in. Allen

and Zies 787

-pots, Causes of failure of. Gorton. . . . 648
Optical, color affected by iron in gases.

Washburn 637

-Tests of sheet and plate. Sorey 801

llaze reduction. Some phenomena in.

Henderson 148

for chemical stoneware 267

-formulas for terra cotta 356

-Polychrome, for terra cotta 353

-Cobalt-uranium green glaze for terra

cotta Wilson 238

for terra cotta, using soluble salts.

Larkin 429

Chinese. Laird 675

lazing. Newcomb Pottery. 524

lycerine, used with underglaze colors 354

Gooch and Browning method for determina-
tion of arsenic 745

Gorton, A. A. "Observations of Apparent
Causes of Failure of Lead Glass

Pots" 648

Granite in Northern Idaho 94

Gravitation, effect upon drying of ceramic

ware. Washburn 25

Greenwood, W. W. Discussion of electro-
lytes in clay 214

Green color for terra cotta glazes. Wilson . 238

soluble, Blends for, on zinc glazes.

Larkin 433

of optical glass 637

Grey color, Soluble, blends for terra cotta.

, ■- Larkin 435

'•^Jrog, in terra cotta bodies. Minton 185

in chemical stoneware 267

in glass pot mixtures 17

Effect of, on strength of saggers 722

Grinding machine for enamel flits 540

Ground coats, Application of, on cast iron . . 543

enamels for cast iron. Staley 99

Gypsum, Hydration of anhydrite and dead-
burned 65

Effect of addition to hydrated lime. ... 175

Calcined, Time of set of. Householder 578

H

Harrop, C. B. "Test of a Producer-gas

Fired Periodic Kiln" 35

Discussion of draft movements in flues. 308

Discussion of heat balance on con-
tinuous kiln 334

Hard porcelain, Some types of 606

fire porcelain, Special 697

Heat distribution in producer-gas fired

periodic kiln 60

balance of chamber continuous kiln.

Treischel and Robertson 322

of dissociation of clay 574

Henderson, H. B. "Some Phenomena in

Glaze Reduction" 148

Hice, R. R. Discussion of Quebec kaolin. . 13

Hill, E. C, quoted on failure of terra cotta. . 185

Discussion of failure of terra cotta

bodies 194

Hillebrand, W. F. Analysis of gas in

bubbles in glass 135

History and composition of alabaster glass.

Silverman 247

of natural hydraulic cements in N. Y.

Jones 160

Holes in glass pots. Causes of 656

Householder. F. F "Some Factors In-
fluencing the Time of Set of

Calcined Gypsum" 578

Howe, R. M. Discussion of tests for silica

brick 345, 499

Hornblende in Idaho granite 96

Hubbard, Prevost. Committee on Cements.. 5

Humidity, Relative, in dryer 32

of atmosphere during heat balance. ... 50

in terra cotta dryers 666

Hursh, R. K. "Heat Balance on a Pro-
ducer-gas Fired Chamber Kiln"... 567
Hydration of anhydrite and dead-burned

gypsum. Gill 65

Hydrated lime pastes, Tests of 173

Hydraulic cements in New York. Jones. .. 160

mining, of Fla. clays 319

properties of the calcium aluminates.

Bates 679

888

JOURNAL OF THE

I

[daho, I'm, clays in northern Sop« 94

IlUnols bond days, Occurrence of ... *W

Impact tests on sheet and plate glass «"

Imported clays, Value <>f Z\7
Impurities in fused magnesia
[^"of refraction, Relation of. to brUhancy

of enamels. • • • • ■ „ .

Industrial Furnace Section, U. S. huci

Administration ••-.•• .,,.

Inspection, visual, of silica brick "V

Iron. Effect of, in silica brick »«

coloring power in glass... .. . . . . •

Effect of. on color of salt glazes. "g-?

Removal of. from zirconium oxide /vi

in glass, Effect of, on pots "^

Removal of. from glass batches . . 652

in K!ass furnace atmosphere. Wash-

hum • '.' ' -'- ,

in optical glass, Determination of "'

J

Jackson. C. E. "A Conveyor Stove ^

Room" .■■■'.•''■ ' V

Jena optical glass, Determination of ^

arsenic in .... 7g4

glasses, Analyses of 529

Jiggering, Ppttery . _ Hydraulic

Tones, Robert W. wa*ur<V,, f i0o

Cements in New York' >™

K

Q

Kaolin in Quebec

Formation of ' •'

in Idaho clay deposits. *°

Occurrence of, in Florida Jl£

American, in porcelain bodies^. 60H

Occurrence of, in America. Ries 440

Constitution of, and rational analyses . . 405

Use of, in Chinese pottery - . . • 6/H

Use of, in coating glass furnace walls. 639

Effect of, on viscosity of enamels... 04,;

Keele, J. "Kaolin in Quebec" 8

Kerr, C. H. Discussion of pots for optical

glass • ■ ■. ■

Discussion of molecular compounds in ^

glass : . c/r

Kentucky bond clays, Occurrence of ...... . 430

Keuffel, C. W. Discussion of seeds in

optical glass • - ■ ■

Kiln, periodic, Description of, used in heat

balance .,.

for burning natural cement '°*

Zwermann tunnel. Description of ^

continuous. Heat balance of -\\-

continuous, Use of draft gages on 3 O

down draft, Use of draft gages on.. . . -MO

producer-gas fired, Heat balance on. 56/

at Newcomb Pottery 3/°

tunnel, Use of, for heavy clay products.

Lovejoy 628

Kirkpatrick, F. A. and Orange, W. B.
"Tests of Clays and Limes by the

Bureau of Standards Plasticimeetr" 1 70

Discussion of hydration of anhydrite . 6V

Kittanning, Lower, clay • ■ • • 2/

Klein A. A. Discussion of tests for silica

brick • • 345

Kramm, H. E. "Work on Gypsum and

Anhydrite o;>

L

Laird, I S "Composition ol ( hi

i h.wIom Pottei v" 61 s

Lake County, Plorida, claj * ' **

Landrum, l< l> Discussion of ground coat

i namels ' ' '

Larldn Paul G "Colors Produced i>\ the

I of Soluble Metallic Salt
Lava Hows in Idaho TJ

Lead and barium, Determination "• in glass 777

in till enamels for east iron *0I

glass, Pot foi Gorton 648

I.eadless tin enamels for cai t iron 704

Leaks in glass pots. Discussion of 651

Lenses from MgO crystals 733

Levol's Method for determination of

arsenic '44

Lime. Tests of. with plastieimcter 170

Effect of. in silica brick 342

Effect of addition of, to porcelain

bodies r,2£

Use of, in silica biick • • I

Use of, in enameling furnaces 645

Lime-alumina-silica system. Compounds of. 679

Limestones, Classification of, after burning 160

in Florida 318

Linings foi glass pots. Discussion of . . ..... 657

Linbarger, S. C. Discussion of effect of

gravity in drying 31

Load test, on silica brick 49*

for silica brick 341

Discussion of. for fire brick 3B

LOCAL SECTIONS:

Northern Ohio Section, June 10, 1918 . 7i

Northern Ohio Section, Dec, 1918. 581

Pittsburgh District Section, December,

1918 lM

New England Section, January,

1919 ■•• ■ 73>

Loomis, George A. "Porosity and Volume
Changes of Clay Fire Brick at

Furnace Temperatures" 38'

Lovejoy, Ellis. Discussion of glaze re-
duction ; ■ ' 5I

Discussion of draft movements in

flues •>«

Discussion of air in clay '*

Discussion of terra cotta dryers ........ 60

"Use of Car Tunnel Kilns for Brick

and Other Products of Crude

Clays" ••-. 62

Lower Kittanning clay, Occurrence of. in

Ohio ••;■ "6

Lower Mercer Clay, Occurrence of, in

Ohio n

M

Machine, for breaking hot briquettes ..
Magnesia, Effect of addition of, to porcelain

bodies Tj

ware, Manufacture of

Sintering of, Ferguson *-

ware. Yensen ; ■ ■ • • • • ■ ■ • ; •

Manometer used in measuring draft in

flues 7;

Manufacture of Alabaster glass £

of silica brick, Methods of

Manufactured products, table, comparative |
Map, showing location of Fla. clay mines ■

Masques used in polychrome decoration 5.

Melting of enamel frits. Poste and Rice /.

enamels for cast iron.

Melts, optical glass, Seeds in

AMERICAN CERAMIC SOCIETY.

889

Metallic cobalt in enamels 112

Metamorphosis of sandstones 479

McDougal, Taine. Discussion of specific

gravity test for porcelains 2H'>
McDanel, W. W. and Riddle, F. H. "Some

Types of Porcelain" 606

Micas in clays and rational analysis 405

in Idaho granite 96

Micro-photographs of crystalline salt

glazes 150

Minton, R. H. Discussion of heat balance

on pel iodic kiln 63

"Use of Furnace Slag as Grog in

Architectural Terra Cotta Bodies" 185

Discussion of green terra cotta glazes. . 244

Discussion of glaze reduction 157

Discussion of slag in terra cotta bodies 194

Discussion of terra cotta glazes 366

Mining hydraulic limestones 166

1 Florida clays 319

Minerals in clays and rational analysis. . . . 405

Mineral products, Statistics of 375

Mineralogical composition of clays 411

Milkiness in optical glass. Gelstharp 559

. in optical glass Fenner and Ferguson 468

Mississippi clays. Occurrence of 453

Missouri white clays. Occurrence of 464

f flint clay areas 465

Mixing of raw materials for cast iron

enamels 534

Modulus of rupture of porcelain bodies. . 610

I of rupture of silica brick 491

"Modal" composition of clays 411

Molding, Effect of electrolytes on 205

Moisture. Determination of, in glass

samples 780

KfFeet of, on dry strengths of clays ... 114

Moist closet for storing clays 29

Mold for casting glass pots 19

Molecular compounds in glass. Tillotson.. 76
Montgomery, E. T. Discussion of heat

balance on periodic kiln 63

Montgomery, R. J. Discussion of tests for

silica brick 350

— • — and Office, L. R. "Notes on the Lab-
oratory Testing of Silica Brick". . 338
Moore, H. W. Discussion of slag in terra

cotta bodies 194

Moscow, Idaho, Clays in vicinity of 94

N

Natural gas in burning salt glazed ware . 4

—hydraulic cements in N. Y. Jones,. 157

Neat portland cements. Discussion of 693

Newcomb Pottery, History and develop-
ment of. Sheerer and Cox 518

'Norms," calculated from analyses of

clays. Ries 418

North Carolina kaolins, Occurrence of 451

o

Ohio stoneware clays, Properties of.

vSchurecht 267

Oil-fired kiln, Zwermann tunnel 262

Opacifier, Antimony oxide used as an 502

Opaque glasses, History and composition of.

Silverman 247

Opalescence in optical glass. Gelstharp . . . 559

-in optical glass 468

Opal glass, Use of fluoride in 561

Optical glass, Special pots for the melting

of. Bleininger 15

-Seeds in. Williams 134

Optical methods of analysis of boric acid and
oxides of arsenic in. Allen and

Zies 739

R61e of arsenic in. Allen and Zies. . . . 787

Color affected by iron in gases. Wash-
burn. 637

Milkiness in Gelstharp 559

Effect of impurities in causing milkiness

in. Fenner and Ferguson. . . 468, 559

Analyses of 784

properties of crystals in glass 603

Opal glass, Use of fluorine in 561

Orton, Edward, Jr., Biography 6

Ortman, F. B. Discussion of heat balance

on periodic kiln 64

Orthosilicate in portland cement 687

Osmotic pressure, Measurement of, for

dilute solution 29

Overglaze, Spraying, on terra cotta 354

Over-burning of fire-bricks, Discussion of. . 395

Oxidation of clays in car tunnel kilns 633

Oxide, zirconium, Purification of. Phillips 791

of arsenic, Methods of analysis of.

Allen and Zies 739

P

Painting terra cotta glazes 353, 429

Paper clays, Occurrence of, in America 447

Parmelee, C. W. Discussion of glaze re-
duction 156

"Unusual Cause of Spalling of Sewer

Pipe" 660

Pastes, Preparation of lime 173

Paving block burning in continuous kiln . . 323

Pearl ash. Analysis of 249

Pegmatite dikes in Idaho 96

Pennsylvania stoneware clays 267

kaolins, Occurrence of 461

Pentavalent arsenic in glass 754

Peregrine, C. R. Discussion of glass pots. . 24

Periclase, Formation of 440

Periodic kiln, Test of producer-gas fired. . . 35
Phillips, A. J. "Partial Purification of

Zr():" 791

Piedmont plateau. Clays in 461

Pitot tube for me isuring draft 296

Pittsburgh Plate Glass Co., Arsenic in

glass of 789

Plasticimeter, Bureau of Standards 170

Plasticity of clays and limes, Measure-
ment of 170

"Plasticity figure" for lime pastes 176

Plastic kaolins of Florida / 317

Plaster, Factors influencing time of set of.. . 578

of paris, Rffect of, on setting of cement 689

Plate glass. Arsenic in 787

Tests of. Sorey 801

Platinum, Effect of, in determining arsenic

in glass 757

Polychrome decoration of terra cotta. . 429, 353
Porosities, bonding clays, at different

temperatures 123

of clays. Effect of electrolytes on 209

of silica brick 343, 489

changes of fire brick at high tempera-
tures 384

Pore water in terms of clay volume 609

Porcelain body slips, Viscosity of 179

bodies, Formulas of 180

used as grog in terra cotta 189

Determining the structure of 287

Some types of 606

Characteristics of 697

grog. Use of, in glass pot mixtures. . . . 657

890

JOURNAL OF THE

Port 1 lain 1 >< ►< 1 %■ containing »<> fret lilii a

bod] . cone 11 623

Portland cement, Compounds in 679
1 p 1 >r 1 u/ 1. hi Hi {i uunil coat

enamels ill

— and Rice, B. A. "Effect of the Degree
of Smelting upon the Properties

of ■ Frit" 221

Discussion of antimony oxide enamels. .Si 1
Discussion of acid test for enamels, 425

Discussion of luster of enamels 646

l'ot mixtures for optical glass pots 16

l'ots, Cooling of optical glass 134

is. Causes of failure of 648

Pot stones in glass 596

Pottery, History of Newcomb 518

clays in America, ( Occurrence of 446

Conveyor stove room for 529

Chinese Celadon. Laird 675

Burning of, in car tunnel kilns 628

Potash, Effect of, in causing milkiness 562

used in optical glass melts . 472

Producer-pas fired periodic kiln, Test of.

Harrop 35

firing of continuous kiln 323

Production of natural cements in N. Y 169

Pressure ratio of gases in flues 303

Protective coatings for glass 801

Pseudowollastonite in glass 601

Pugged clays. Air content of 584

Pvcnometer-volumeter. Schurecht 556

Q

Quaxtzite, Glenville, Description of 8

Physical properties of 479

Quartz schist, in northern Idaho 94

Properties of 477

in clay, Rational analysis of 407

Effect of, in porcelain bodies 697

Quarries, Limestone, in New York 160

Quicklime. Use of, in enameling furnaces... 645

R

Radcliffe, B. S. Discussion of terra cotta

glazes 366

Rand, C. C. Discussion of compounds in

glass 90

Rational analysis of clays. Washington.. 405

Reducing conditions in gas fired kiln 62

Refractive index, glass 78

Refractoriness of Idaho fire clay 97

of silica brick 343

Measurement of 385

of saggers 719

Refractory flint clays of Mo., Occurrence of 466
Refractories, Volume changes of, at high

temperatures 384

Zr02, Purification of. Phillips 791

Residual clays in northern Idaho 94

Retorts for zirconia purification 794

Rice, B. A. and Poste, E. P. Enamel frits . . 221

Discussion of acid test for enamels. . . . 426

Riddle, F. H. Discussion of gravity drying 32

and McDanel, W. W. "Some Types of

Porcelain" 606

Ries, H. Discussion of Quebec kaolin 12

"The Occurrence of High-grade Amer-
ican Clays and Their Further

Development" 446

Robertson, H. S. and Treischel, C. "Heat

Balance of Continuous Kiln" 322

Roman cement, Production of, in N. Y 160

Rosendale cement, in N. Y 160

Ross, D. W. "Silica Refractories" 477

Discussion of tests for silica brick . . 349, 500

Ro i> w. Discussion ol pots foi optical

glass 2<

KussianpolasliiniiplM.il glass 469,561

s

Sagger grog in terra COttS bodies 18'

clays and mixtures. Brown 7I(

Salts, soluble, in terra cotta glazes 353, 42'

Salt-water, Formation Of, On optical glass 47.

Sand carrying capacity of limes 17.

Sands, Kaolin bearing, of Pla 311

Saxe, C. W. and Buckner, O. S. "Bonding

Strengths of Clays" 11;

Schurecht, II (', "Effect of Electrolytes

on Properties of Clays" 20

"Properties of Some Ohio and Penna.

Stoneware Clays" 26'

"Pycnomcter Operated as a Volumeter" 55i

Screens for sieving enamel frits 54

Seeds in optical glass melts. Williams 13'

in glass. 59'

Seger volumeter, Discussion of 55<

Sellards, E. H. "The Clays of Florida" U.

Semi-flint clay. Occurrence of, in Ohio 27!

Setting of natural anhydrite and dead-
burned gypsum 6V

of calcined gypsum 571

properties of tri-caleium aluminate. ... 68

Sewer pipe, Crystals in salt glazes of 14'

Causes of spalling of 661

Shaw, J. 1!. "Antimony Oxide as an Opaci-

fier in Cast Iron Enamels" 50;

Discussion of electrolytes in clay 21'

Discussion of smelting of enamel frits.. 23.

Discussion of acid test 42

Sheet glass, Tests of. Sorey 80

Sheerer, M. G. "Newcomb Pottery" 5!

Sheppard, E. S. Determination of gases in

glass 7ft

Shrinkage of terra cotta bodies 19

volume, Effect of electrolytes on 20

Sieve analysis of sagger grog 72

Sifters for powdered enamels 54

Silica brick, from Quebec quartzite 1

washed from Idaho fire clay 9

brick. Testing of 33

Effect of, in alabaster glass 25

bricks in glass furnace crowns 59

Separation of, from optical glass 55

Use of, in cast iron enamels 50

refractories. Ross 47

content of sintered magnesia 44

LTse of, in sagger mixtures 72

Effect of, on luster of enamels 64

Silicates, Determination of boric acid in.. 76

Silicious clays in sagger mixtures 72

Sillimanite, Use of, in hard porcelain 62

Formation of, in glass pots 65

Effect of, in porcelain bodies 69

in glass pot stones 59

Sintering of magnesia. Ferguson 43

Sintered ground coat enamels for cast iron . 9
Silverman, A. "Alabaster Glass; History

and Composition" 24

Slag, Use of, in terra cotta bodies 18

Slip, Clay, for glass pot casting

for terra cotta. Larkin 43

terra cotta, Composition for 35

Slaking of clay, Determination of structure

by 71

Soaking pit for use in preparing clay 72

Sodium sulphate freezing tests on terra

cotta 1*

Soda-lime glass, Opalescence in 56

AMERICAN CERAMIC SOCIETY.

891

Sodium metantimonate enamels 706

sulphate scum on enamels 644

Softening points of Ohio and Penna

stoneware clays 279

of silica brick 338

of kaolin-feldspar-quartz mixtures 616

Determination of, for fire brick 394

Discussion of, foi file brick 397

Soluble salts. Use of, in terra cot to glazes. . 353

Solubility of enamels for cast iron 506

Solutions, Standard, for soluble colors 432

Acetic acid, for testing enamels 422

Soper, E. K. "Fire Clays in Northern

Idaho" 94

Sorey, T. I.. "Tests of Plain and Protec-
tive Sheet Glass" 801

South Carolina kaolins. Occurrence of 456

Spark plug porcelains 699

Specific gravity determination on glass 90

of silica brick . 342

test for porcelains 287

test for silica brick 351, 485

Specifications, load test on silica brick 495

for fire brick 389

Spencer I.ens Co., Arsenic in glasses of 789

Spraying of terra cotta glazes 353, 429

Spuirier, H. "Use of Dense Solutions in
Determining the Structure of

Porcelain" 287

Discussion of sp. gr. tests for porce-
lains 289

"Nature of the Air Content of

Pugged Clays" 584

"Methods for the Determination of

Air in Plastic Clays" 710, 715

Staley, H. F. "Ground Coat Enamels for

Cast Iron" 99, 111

"Preparation and Application of

Enamels for Cast Iron" 534

"Enamels for Cast Iron" 703

"The Control of the Luster of

Enamels" 640

Discussion of glaze reduction 159

Discussion of electrolytes in clays 214

— — -Discussion of SB2O3 enamels 511

Discussion of enamel frits 235

Stains, green, for terra cotta 238

Static pressure of gases in flues 294

Stack, Pressure of gases in 294

Still, Platinum, used in determination of

arsenic 758

Stirring cast iron enamels 539

optical glass melts 134

of gypsum pastes. Effect of 581

Stoneware clays of Ohio and Penna 267

Strength of special portland cements 689

Strength of terra cotta bodies 192

of clays, Effect of electrolytes on 201

Crushing, of silica brick 492

Storage of enamel raw materials 534

Effect of, on cement and concrete 694

Stones in glass, Source of 654

Identification of. in glass 594

Stove room with conveyor 529

Striae in glass 594

Stnll, R. T. Discussion of glaze reduc-
tion 158

Sulphur compounds, Effects of, on enamels. 644

Sulphur dioxide in glass bubbles 139

in glass, Determination of 782

Sulphates in optical glass as a source of

milkiness 56 1 , 473

Summer meeting, ACS 141

Swalm 1'. H. Discussion of slag in terra

cotta bodies 193

St. Remi, Quebec, kaolin 8

T

Table, enameling, Description of 550

Tails in glass 597

Tannic acid, Effect of. in clays 203

Temperature-, Effect of, on gravity drying. 28

Effect of, on set of gypsum 583

in enameling furmces 554

control in glass melting 471

in fining optical ulass 131

distribution in car tunnel kilns 633

Tensile strengths of cement mortars 690

Tennessee bond clays, Occurrence of 456

Terra cotta bodies, Use of slag in 185

glaze, Cobalt -uranium •. . 238

Polychiome decoration of 353

glazes, Use of salts in 429

dryers 667

Testing silica brick 338

of porcelains by solutions 287

Tillotson, E. W. Discussion of pots for

optical glass 23

"Relation between Physical Proper-
ties and Chemical Composition of

Glass" 76, 91

Tillyer, E. D. Discussion of compounds in

glass 89

Tin enamels, Fritting 539

for cast iron. Stalev 703

Tin oxide in enamels, Substitute for 502

Tionesta clays. Occurrence ot, in Ohio 269

Treischel, C. and Robertson, H. S. "An
Attempted Heat Balance on a

Continuous Kiln" 322

Discussion of heat balance on con-
tinuous kiln 336

Tricalcium aluminate, Properties of 680

Tridymite, Optical properties of 604

in silica brick 488

Truman, Gail. Discussion of terra cotta

glazes 365

Tunnel kiln, Operation of. Zwermann. . . . 262

Use of, for heavy clay products 628

Tunnel dryers for terra cotta 666

u

Underglaze colors. Formulas of 354

Uranium-cobalt glaze for terra cotta 238

I S Brick Co., Tests of periodic kilns at.. 35

U. S Fuel Administration 2

V

Vapor pressure, Influence of, in drying 27

Velocity of gases in flues 294

Vicat apparatus, Use of, in measuring set

of plaster 578

Viscosity, of optical glass 135

of clays and porcelain body slips 179

Measurement of, of gypsum pastes. . . . 580

of enamels, Effect of, on crystalliza-
tion 640

Virginia clays, Occurrence of 462

Volatilization, of glass, Effect of, on index

of refraction 90

from enamel frits 221

in optical glass melts 560

of arsenic from glass 741

-of iron chloride 799

I' »i RNAL <)K THE

\ ol< anii i nil . in Idaho 9 i

Volurai ter, Pycnometei used i 536

W

U allastonite in glass 601
\\ irpage "i Ohio and Penna stoneware

clays 279

Washburn, E W. Committee on ceramics 5
"Effect «f Gravitation upon the Dry-
ing of Ceramic Ware" 25

Discussion of glaze reduction 158

Presence of iron in the Furnace
Atmosphere as :i Source of Color
in tho Manufacture of Optical

Glass" ... 637

Washing, Idaho fire clays 97

Ohio and Penna. stoneware clays 273

Florida kaolins 'IK

Georgia kaolins 451

Washington, II. S. ■Calculation of the

Rational Analysis of Clays" 405

Water glass, Stability of 86

— ■ — glass as binder for carborundum 268

of plasticity, Effect of electrolytes on. . 205

Waste heat for terra cotta dryers 666

Watts, A. S. Discussion of gravity in

drying -. 31

Discussion of sp. gr. test lor porcelain.. 290

Wei pan, i m ol Foi taggi i ciaj 728

win i r ' - method id di tei initiation ol Bt< I

in glai

Whiiakir, K A Discussion ol electro

in clay

Williams, a. I') Discussion ol compo ition

of glass, , , 90

"Observations on the Formation <>i

Seed in Optical Class Melts" 134, 137

Wilson, Hewitt, "A Cobalt Uranium

Green Glaze for Terra Cotta". 238, 244
Discussion of terra cotta glazes 366

"Polychrome Decoration of Terra ( otta

with Soluble Metallic Ss 353

Windsor, N. S., Anhydrite from 65

Y

Yellow colors, Soluble blends for terra

cotta. I.arkin 436

Yensen. Trygve 1), "Magtiesia Ware" , .. 730

Z

Zinc oxide, Effect of, in antimony enamels 707

in optical tflass, Determination of 712

Zirconium oxide, Purification of. Phillips.. 791
Zwermann, Carl H. "The Zwermann

Tunnel Kiln and Its Operation". . 262

^

ERRATA SLIP

"Some Phenomena in Glaze Reduction," by H. B. Henderson, Vol. i,

No. 3, page 148 et seq.

Page 149 line 4 from the top omit the word "somewhat."

Page 149 line 12 from the bottom change the word "graphitic" to

"graphite."

Page 151 line 14 from the top omit the words "iron oxide."

Page 151 line 19 from the top insert the words "some of" after the word

"glaze."

Page 152 line 12 from the top insert the word "ferrous" before the word

"iron."

Page 152 change the footnote to read: "Sosman and Hostetter, Trans.

American Inst. Mining Engineers, Bull. 126."

Page 153 line 4 from the top change "will" to "may."

Page 153 line 23 from the top change "color" to "coloring matter."

Page 155 lines 11 and 12 omit the clause "except by the assumption of

transformation and diffusion effects."

Page 155 line 15 from the top change "but" to "and."

Page 155 line 17 from the top change "probable" to "evident."

Page 155 line 19 from the top insert the clause "if it is iron" after the

word "iron."

BINDING SECT. NOV 1 8 1978

TP

American Ceramic Sc

>ciet

785

Journal

A63
v.l

W

cop. 2

Engineering

PLEASE DO NOT REMOVE
CARDS OR SLIPS FROM THIS POCKET

UNIVERSITY OF TORONTO LIBRARY